地球信息科学学报 ›› 2022, Vol. 24 ›› Issue (1): 100-113.doi: 10.12082/dqxxkx.2022.210359

• 地球信息科学理论与方法 • 上一篇    下一篇

耦合FLUS和Markov的快速发展城市土地利用空间格局模拟方法

王旭东1(), 姚尧1,2,*(), 任书良1, 史绪国1   

  1. 1.中国地质大学(武汉)地理与信息工程学院,武汉 430078
    2.阿里巴巴集团,杭州 311121
  • 收稿日期:2021-06-28 修回日期:2021-09-22 出版日期:2022-01-25 发布日期:2022-03-25
  • 通讯作者: * 姚尧(1987—),男,广东梅州人,副教授,研究方向为空间大数据和城市计算。E-mail: yaoy@cug.edu.cn
  • 作者简介:王旭东(1998— ),男,湖北荆州人,硕士生,主要从事地面沉降和城市发展驱动机制研究。E-mail: wxd@cug.edu.cn
  • 基金资助:
    国家自然科学基金项目(41801306);国家重点研发计划项目(2019YFB2102903)

A Coupled FLUS and Markov Approach to Simulate the Spatial Pattern of Land Use in Rapidly Developing Cities

WANG Xudong1(), YAO Yao1,2,*(), REN Shuliang1, SHI Xuguo1   

  1. 1. School of Geography and Information Engineering, China University of Geosciences, Wuhan 430078, China
    2. Alibaba Group, Hangzhou 311121, China
  • Received:2021-06-28 Revised:2021-09-22 Online:2022-01-25 Published:2022-03-25
  • Supported by:
    National Natural Science Foundation of China(41801306);National Key Research and Development Program of China(2019YFB2102903)

摘要:

模拟城市土地利用空间变化格局的研究,对未来区域规划以及实现可持续发展具有十分积极的作用。以往基于FLUS的研究栅格尺度较大,如何模拟快速发展中城市的复杂土地利用变化过程,挖掘土地利用变化驱动机制值得进一步探讨。本文构建了耦合FLUS和Markov的城市土地利用格局拟合框架,创新性地引入房价指标表征社会经济属性,以深圳为研究区,基于30 m空间分辨率小栅格尺度的土地利用分类数据和基础地理、路网河网、感兴趣点等多源空间变量,模拟不同发展情景下的未来城市土地利用空间格局,并通过随机森林进行土地利用变化驱动因素分析。研究结果表明:本文提出的耦合FLUS和Markov方法相较于传统CA模型(RFA-CA和Logistic-CA)精度更高(FoM=0.22),能更准确地模拟快速发展中城市的土地利用变化过程;多情景土地利用格局制图结果验证了城市发展过程中生态控制线的重要性,进一步说明本文拟合框架在未来城市规划布局中的参考价值;医院、娱乐场所等的基础设施和公交、路网密度等的基础交通比自然因素(高程、坡度)对城市发展的影响更大,到海岸线距离会在一定程度上限制深圳内部土地利用变化过程。本研究所构建模型及精细制图结果,可为城市区域规划和空间格局模拟等相关研究提供参考依据和理论基础。

关键词: 多源数据, 快速发展中城市, 土地利用格局, 马尔可夫模型, 未来土地利用模型, 情景设置, 土地利用空间格局模拟, 土地利用变化驱动机制

Abstract:

Modeling urban land use change is important for future regional planning and sustainable development. Previous FLUS-based studies are mostly based on larger grid scales. How to simulate the complex land use change processes in rapidly developing cities and explore the driving mechanisms of land use change still need further exploration. This paper constructs an urban land use pattern simulation framework coupled with FLUS and Markov and innovatively introduces house price to characterize socio-economic attributes. We take Shenzhen as the study area to simulate future urban land use spatial patterns under different development scenarios based on small grid scale (30 m) land use classification data and multi-source spatial variables such as basic geography data, road and river networks, and point-of-interest data. Finally, we analyze the land use change drivers using random forest models. The results show that the coupled FLUS and Markov method proposed in this paper has higher accuracy (FoM = 0.22) and simulate the land use change processes more accurately in rapidly developing cities, compared to traditional CA models (RFA-CA and Logistic-CA). The mapping results of multi-scenario land use patterns verify the importance of ecological control lines in the process of urban development, further illustrating the reference value of the proposed simulation framework for future urban planning layout. Hospital infrastructure, entertainment venues, and bus stop, road network density have a greater impact on urban development than natural factors (e.g., elevation, slope), while the distance to coastline limits land use change processes to some extent within Shenzhen. The model constructed in this study and the fine mapping results can provide a reference basis and theoretical foundation for related research on urban regional planning and spatial pattern simulation.

Key words: multi-source data, rapidly developing cities, land use patterns, Markov model, FLUS model, scenario settings, simulation of land use spatial patterns, land use change driving mechanisms