基于GBDT机器学习算法的国土空间演变动力机制研究——以长江三角洲为例  

Research on Dynamic Mechanism of Territorial Space Evolution in Yangtze River Delta Based on GBDT Machine Learning Algorithm

作　　者：夏陈红 翟国方[1] XIA Chenhong;ZHAI Guofang

机构地区：[1]南京大学建筑与城市规划学院,江苏南京210008

出　　处：《城市发展研究》2022年第11期I0012-I0019,共8页Urban Development Studies

摘　　要：国土是生态文明建设的空间载体,开展国土空间演变的内在规律与动力机制研究,对推动形成完善的城镇化推进格局、农业发展格局与生态安全格局具有重要意义。目前我国对国土空间演变的研究多依赖于传统数理模型,极少将机器学习算法引入具有普适化的区域性研究。本研究以长江三角洲为例,在运用国土空间演变理论与方法刻画1995~2020年长江三角洲国土空间时空序列变化规律的基础上,引入梯度提升决策树(GBDT)机器学习模型揭示多维时空和多元系列下国土空间格局形成的动力驱动机制。结果表明:(1)长江三角洲国土空间结构呈现出明显的城镇空间扩张与农业空间收缩交替、生态空间波动式扩张的现象。(2)长江三角洲城镇空间格局变化速度最快,且逐步由大规模扩张转向精明发展;农业空间格局变化速度相对平缓,但整体收缩态势明显;生态空间格局变化速度较低,空间发展稳定。(3)常住人口是城镇空间格局演变最重要的驱动力,公共财政支出和全社会固定资产投资总额次之,且三者之间具有显著的协同增强作用;农业机械总动力是农业空间格局演变最重要的驱动力,年平均气温和农民人均可支配收入次之;森林面积是生态空间格局演变的主导性驱动力。Territorial space is the spatial carrier of ecological civilization construction.It is of great significance to study the internal laws and dynamic mechanisms of national territory spatial evolution.Taking the Yangtze River Delta (YRD) as an example,this study introduced the gradient lifting decision tree (GBDT) model to reveal the dynamic driving mechanism of territorial spatial pattern formation based on the spatial analysis method to describe the evolution of the territorial spatial pattern in the YRD from 1995 to 2020.The results showed that:(1) The YRD territorial space showed apparent urban space expansion alternating with agricultural space contraction,and ecological space fluctuated.(2) The urban spatial pattern change rate in the YRD was the fastest,gradually turning from large-scale expansion to smart contraction.The agricultural spatial pattern change rate was relatively flat,but the overall contraction trend was apparent.The ecological spatial pattern change rate was relatively low,and the spatial development was stable.(3) Permanent resident population is the most important driving force for the evolution of urban spatial patterns,followed by public financial expenditure and total social investment in fixed assets.The total driving force of agricultural machinery was the most important driving force for the evolution of agricultural spatial patterns,followed by the annual average temperature and per capita disposable income of farmers.Forest area is the dominant driving force of ecological spatial pattern evolution.

关 键 词：国土空间 演变动力机制 梯度提升决策树(GBDT) 长江三角洲 

分 类 号：TU984[建筑科学—城市规划与设计]