摘要
深圳土地开发强度高，导致生产、生活与生态功能在有限空间内交织，功能叠置与空间分异并存。本文以12514个街区单元为对象，利用高德兴趣点（POI）、开放街道地图（OSM）路网和1 m分辨率不透水面数据划分空间类型，并结合自然生态与社会经济变量，采用空间自相关、基尼系数、区位熵、普通最小二乘回归和地理加权回归分析其分异特征与形成条件。结果表明，生产空间与混合三生空间的集聚性最强，生态相关空间相对分散；混合三生空间在纳入统计的9个行政区均有分布，生活空间在福田、罗湖和南山等成熟城区总体占优，在龙岗、龙华等人口密集地区也表现出一定集聚。生产和生活空间主要与经济发展、人口、道路和人类活动相联系，混合三生空间还受到植被覆盖等生态条件影响，生态空间则与植被和地形条件联系更紧密。地理加权回归模型（GWR）在多数空间类型上的拟合度有所提高，表明全局平均关系难以完全解释城市内部差异。深圳空间治理可依据局部集聚类型和总体关联特征识别功能优化重点，并通过动态监测与滚动调整提高治理响应的针对性。
关键词： 三生空间；空间分异；空间自相关；地理加权回归；深圳市
Abstract
Shenzhen is characterized by intensive land development, with production, living, and ecological functions concentrated within a limited urban space, where functional overlap and spatial differentiation coexist. This study examines 12,514 urban blocks. Utilize amap Points of Interest (POI), OpenStreetMap (OSM) road network, and 1 m impervious-surface data are used to classify spatial functions, while natural and socioeconomic variables are incorporated to examine their formation conditions. Spatial autocorrelation, the Gini coefficient, location quotient, ordinary least squares regression, and geographically weighted regression are applied. Production space and mixed production-living-ecological space show the strongest agglomeration, whereas ecological-function spaces are relatively dispersed. Mixed-function spaces occur across the nine districts included in the administrative comparison. Living space is relatively concentrated in mature districts such as Futian, Luohu, and Nanshan, with additional clusters in densely populated Longgang and Longhua. Production and living spaces are closely associated with economic development, population, road conditions, and human activity. Mixed-function space is also related to vegetation cover and other ecological conditions, while ecological space is more strongly associated with vegetation and topography. The Geographically Weighted Regression (GWR) model improves model fit for most spatial types, indicating that citywide average relationships do not fully explain intra-urban variation. Spatial governance in Shenzhen may therefore use local clustering patterns and overall statistical associations to identify functional optimization priorities, supported by dynamic monitoring and iterative adjustment.
Key words： Production-living-ecological space; Spatial differentiation; Spatial autocorrelation; Geographi-cally weighted regression; Shenzhen
参考文献 References
[1] 李广东, 方创琳. 城市生态—生产—生活空间功能定量识别与分析[J]. 地理学报, 2016, 71(1): 49-65.
[2] 江东, 林刚, 付晶莹. “三生空间”统筹的科学基础与优化途径探析[J]. 自然资源学报, 2021, 36(5): 1085-1101.
[3] ZHANG W, YUAN Q, CAI H. Unravelling regional development through the production-living-ecological perspective: Assessing heterogeneity and expert insights[J]. Urban Climate, 2024, 55: 101937.
[4] LI H, ZHANG P, ZHANG L, et al. Research on the evolutionary characteristics and mechanism of production-living-ecological space in Shanxi Province, China[J]. Scientific Reports, 2024, 14: 31379.
[5] 刘继来, 刘彦随, 李裕瑞. 中国“三生空间”分类评价与时空格局分析[J]. 地理学报, 2017, 72(7): 1290-1304.
[6] 孔冬艳, 陈会广, 吴孔森. 中国“三生空间”演变特征、生态环境效应及其影响因素[J]. 自然资源学报, 2021, 36(5): 1116-1135.
[7] FU J, BU Z, JIANG D, et al. Sustainable land use diagnosis based on the perspective of production-living-ecological spaces in China[J]. Land Use Policy, 2022, 122: 106386.
[8] LI S, CHEN Y, HUANG Q. Spatial allocation and sustainable development: A study of production-living-ecological spaces in the Yangtze River Delta[J]. Frontiers in Environmental Science, 2025, 12: 1481186.
[9] WU S, MO W, ZHANG R, et al. Spatiotemporal evolution and driving mechanism of production-living-ecological space from 1990 to 2020 in Hunan, Central China[J]. Sustainability, 2025, 17(4): 1703.
[10] 杨振山, 苏锦华, 杨航, 等. 基于多源数据的城市功能区精细化研究——以北京为例[J]. 地理研究, 2021, 40(2): 477-494.
[11] YUAN N J, ZHENG Y, XIE X, et al. Discovering urban functional zones using latent activity trajectories[J]. IEEE Transactions on Knowledge and Data Engineering, 2015, 27(3): 712-725.
[12] ZHANG X, DU S, WANG Q. Hierarchical semantic cognition for urban functional zones with VHR satellite images and POI data[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2017, 132: 170-184.
[13] ANSELIN L. Local indicators of spatial association—LISA[J]. Geographical Analysis, 1995, 27(2): 93-115.
[14] FOTHERINGHAM A S, BRUNSDON C, CHARLTON M. Geographically Weighted Regression: The Analysis of Spatially Varying Relationships[M]. Chichester: Wiley, 2002.
[15] 张令达, 侯全华, 段亚琼. 生态文明背景下三生空间研究：内涵、进展与对策[J]. 生态学报, 2024, 44(1): 47-59.
[16] 陈阳, 岳文泽, 张亮, 等. 国土空间规划视角下生态空间管制分区的理论思考[J]. 中国土地科学, 2020, 34(8): 1-9.
[17] 林良国, 赵耀龙, 柯恩童. 面向城市更新的地理空间信息技术研究现状与展望[J]. 地球信息科学学报, 2024, 26(4): 898-914.