Objective The spatial pattern and driving factors of landscape fragmentation in Beijing-Tianjin-Hebei region were investigated in order to provide a scientific basis for land-use spatial optimization and sustainable regional development in Beijing-Tianjin-Hebei urban agglomeration. Methods Taking Beijing-Tianjin-Hebei region as a study area, three time points (2000, 2010, and 2022) were selected, and nine natural and human factors were introduced. Spatial autocorrelation, the XGBoost-SHAP model, and the PLS-SEM model were integrated for comprehensive analysis. Results ① Landscape fragmentation in Beijing-Tianjin-Hebei region generally showed an intensifying trend and exhibited significant spatial clustering effects. High-fragmentation hotspots were mainly concentrated in the northwestern Yanshan Mountains and along the Taihang Mountains. ② The XGBoost-SHAP model revealed that the importance ranking of the driving factors of landscape fragmentation was as follows: land cover change intensity > slope > soil organic carbon content > human activity intensity > elevation > annual precipitation > annual average temperature > impervious surface expansion intensity > human footprint. Among these factors, land cover change intensity, slope, annual precipitation, and annual average temperature primarily exerted positive driving effects on landscape fragmentation, whereas soil organic carbon content and human activity intensity showed negative inhibitory effects. ③ The PLS-SEM path analysis further clarified the action pathways of each factor and their interactions. Land cover change intensity had a direct positive effect on landscape fragmentation, whereas slope, soil organic carbon content, elevation, annual precipitation, annual average temperature, and impervious surface expansion intensity mainly exerted indirect effects by influencing human activity intensity, land cover change intensity, and human footprint. Conclusion Over the past 23 years, the degree of landscape fragmentation in Beijing-Tianjin-Hebei region has intensified. Land cover change intensity and slope are the main driving factors of landscape fragmentation. Land cover change intensity directly intensifies landscape fragmentation, whereas the other factors mainly influence landscape fragmentation indirectly by affecting human activities.
文献参数: 王贺封, 姜涵, 曹媛, 等.基于XGBoost-SHAP和PLS-SEM的京津冀景观破碎化驱动因素研究[J].水土保持通报,2026,46(3):149-159. Citation:Wang Hefeng, Jiang Han, Cao Yuan, et al. Driving factors of landscape fragmentation in Beijing-Tianjin-Hebei region based on XGBoost-SHAP and PLS-SEM [J]. Bulletin of Soil and Water Conservation,2026,46(3):149-159.
偏最小二乘结构方程模型(PLS-SEM)是一种基于方差的多变量分析方法,适用于探究变量及潜变量之间的复杂因果路径与影响机制[26]。该方法不预设严格的数据分布前提,具备处理复杂模型的稳健性,且在潜变量设置方面对显变量无严格的数量要求,并能同步估计变量间的直接与间接效应,从而系统地揭示其内在作用关系。在本文中,共设置两个潜变量,其中气候特征潜变量(Climate)由AP和AT两个特征变量组成,地形特征潜变量(Topography)由DEM和Slope两个特征变量组成。通过拟合优度(Goodness of Fit, GOF)指标评估模型的拟合效果,根据相关研究经验,当GOF大于0.5时,认为结果可靠。相关计算通过R语言实现。
WangXianli, BuRencang, HuYuanman, et al. Analysis on landscape fragment of Liaohe delta wetlands [J]. Chinese Journal of Applied Ecology, 1996,7(3):299-304.
[3]
WuJianguo. Effects of changing scale on landscape pattern analysis: Scaling relations [J]. Landscape Ecology, 2004,19(2):125-138.
LiuJiyuan, KuangWenhui, ZhangZengxiang, et al. Spatiotemporal characteristics,patterns and causes of land use changes in China since the late 1980s [J]. Acta Geographica Sinica, 2014,69(1):3-14.
[6]
FoelskeL, van RiperC J. Assessing spatial preference heterogeneity in a mixed-use landscape [J]. Applied Geography, 2020,125:102355.
WuJianda, ZhangAnbing, WangHefeng, et al. Dynamic monitoring and influencing factor analysis of ecological environment in mining cities [J]. Journal of Northwest Forestry University, 2025,40(4):201-212.
BoLiming, YinLi, WeiWei, et al. Evolution characteristics and influencing mechanism of aquatic ecological spatial pattern of urban agglomeration in the middle reaches of the Yangtze River [J]. Resources and Environment in the Yangtze Basin, 2025,34(7):1441-1455.
ZhangRenfei, XiaoMeng, LiuZhicheng. Spatio-temporal heterogeneity and driving factors of landscape fragmentation in Beijing-Tianjin-Hebei region [J]. Ecology and Environmental Sciences, 2025,34(3):461-473.
LuXiangmei, PengShuangyun, CaiFucheng, et al. Impacts of human activities and climate change on wetland landscape pattern of the nine major plateau lake basins in Yunnan Province [J]. Wetland Science, 2024,22(2):191-206.
WangQuanquan, WangHang, ZhangWeiguo, et al. The correlations between wetland landscape and social-natural factors on northwestern Yunnan Plateau [J]. Acta Ecologica Sinica, 2019,39(2):726-738.
QuNan, LuJie, WangZiming, et al. Spatio-temporal evolution and driving factors of ecological environment quality in Mu Us sandy land [J]. Bulletin of Soil and Water Conservation, 2025,45(5):336-346.
[19]
ChenTianqi, GuestrinC. XGBoost: a scalable tree boosting system [C]∥ Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. San Francisco California USA. ACM, 2016:785-794.
[20]
LundbergS M, LeeS I. A unified approach to interpreting model predictions [C]∥ Neural Information Processing Systems, 2017.
[21]
LefcheckJ S. piecewiseSEM: Piecewise structural equation modelling in R for ecology, evolution, and systematics [J]. Methods in Ecology and Evolution, 2016,7(5):573-579.
ZhangYongbin, LiChunyu, LiuMingyue, et al. Coupling analysis of urbanization and ecological environment in Beijing-Tianjin-Hebei urban agglomeration from 2000 to 2020 [J]. Journal of Applied Sciences, 2025,43(3):475-490.
ZhangYachao, XuLei, ZhangGuijun, et al. Structural characteristics and driving factors of multifunctional symbiosis network of land use in the Beijing-Tianjin-Hebei urban agglomeration [J]. Geography and Geo-Information Science, 2025,41(3):101-110.
WangFuwu, ZhangYuhong. Landscape fragmentation analysis and eco-environmental quality assessment at the catchment scale [J]. Journal of Natural Disasters, 2023,32(1):67-75.
LiangJiale, ChenWanxu, LiJiangfeng, et al. Spatiotemporal patterns of landscape fragmentation and causes in the Yellow River basin [J]. Acta Ecologica Sinica, 2022,42(5):1993-2009.
YanEnping, LinHui, WangGuangxing, et al. Analysis of evolution and driving force of ecosystem service values in the Three Gorges Reservoir region during 1990—2011 [J]. Acta Ecologica Sinica, 2014,34(20):5962-5973.
[34]
JiangSimin, FengFei, ZhangXinna, et al. Driving factors of fragmentation in urban landscapes:Local contributions, spatial relationships, and causal effects [J]. Ecological Indicators, 2025,174:113454.
ZhaoHui, LiuQian, ZhangMin, et al. Analysis of spatial patterns and driving factors of ecosystem services in Beijing based on XGBoost-SHAP model [J]. Environmental Science, 2026,47(2):1025-1037.
TanJie, WeiQianjun, LiaoZhaoyang, et al. Relationship between urban form and surface temperature based on XGBoost SHAP interpretable machine learning model [J]. Chinese Journal of Applied Ecology, 2025,36(3):659-670.
YuanHao, TangDiwei, LiuHeng, et al. Dynamics evolution and driving factors interaction effects of ecosystem services in the Wuling Mountain area [J]. Environmental Science, 2026,47(1):431-443.
ZhangXinsheng, LiYinan, ChenZhangzheng, et al. Research on carbon emission prediction of manufacturing industry in northwest China based on GBDT-XGBoost [J]. Environmental Science & Technology, 2024, 47(12):131-143.
LiuDongyue, DongWenzhuo, GouRong, et al. Spatial and temporal evolution of ecosystem service supply and demand and analysis of driving factors in Chongqing based on InVEST model and XGBoost-SHAP[J/OL]. Environmental Science, 2025,1-17.
LiuYiran, AnFengjun. A comparative study of canonical correlation analysis and structural equation modeling [J]. Statistics and Decision, 2024,40(10):40-45.
JiZhengxin, XuYueqing, HuangAn, et al. Spatial pattern and evolution characteristics of the production-LivingEcological space in the mountainous area of northern Hebei Province: A case study of Zhangjiakou City [J]. Acta Scientiarum Naturalium Universitatis Pekinensis, 2022,58(1):123-134.
MaYinqiu, LiJiahui, CaoWei, et al. Projecting the carbon sink potential and contribution of Grain for Green Program in the Beijing-Tianjin-Hebei region [J]. Acta Geographica Sinica, 2024,79(3):732-746.
ZhuLinna, ZhaoMudan, LiYunfei, et al. The space-time relationship between the ecosystem service value and the human activity intensity in Xi’an metropolitan area [J]. Journal of Ecology and Rural Environment, 2024,40(3):325-334.
[53]
DoetterlS, BerheA A, HeckmanK, et al. A landscape-scale view of soil organic matter dynamics [J]. Nature Reviews Earth & Environment, 2025,6(1):67-81.