| 摘要: |
| 城市蓝绿空间的科学规划有利于缓减
城市热岛、降低能耗和增进城市人群健康。然
而，中国大城市滨水蓝绿空间冷岛效应却缺乏
较为深入的定量分析和研究。本研究以武汉、
南京和杭州三个水系发达城市为例，在借助多
源遥感和GIS空间数据计算三个中国城市地表
温度（LST）及冷岛强度基础上，采用基于机器
学习的决策树回归方法探究各滨水蓝绿空间形
态因子的对城市冷岛效应的影响程度和分布特
征，同时有效克服了传统线性回归模型在揭示
蓝绿空间众多影响因子交互作用方面的受限问
题。结果表明，武汉滨水蓝绿空间的强冷岛效
应区域位于东湖、汤逊湖等大型湖泊周边，南京
则为秦淮河南岸，杭州则为钱塘江中段南侧和
西湖南侧。决策树回归结果表明，不同城市起
主导作用的景观形态因子也存在差异，不同城
市的水域形态对城市热岛影响作用明显。在杭州和南京，水面率和绿地率为主导因子，杭州受其共同影响占72.9%，南京则为61.8%。在武汉，
水面率、水域形状指数为主导因子，受其共同影响占63.5%。研究结果可以为中国大城市滨水蓝
绿空间的科学规划及设计提供有益参考。 |
| 关键词: 城市冷岛强度 蓝绿空间 城市水网形态 决策树回归 |
| DOI：10.13791/j.cnki.hsfwest.20230604 |
| 分类号: |
| 基金项目:国家自然科学基金项目（51578482）；中国气象局气
候资源经济转化重点开放实验室开放研究课题（20
23-15） |
|
| Influencing Factors and Distribution Characteristics of Cold Island Effect in Urban Waterfront Blue-Green Spaces: The Case Studies of Wuhan, Nanjing and Hangzhou |
|
WANG Weiwu,LIANG Shuang,YANG Hanzi
|
| Abstract: |
| The process of urbanization has led to the replacement of natural landscapes, such as
vegetation and water bodies, by man-made structures, triggering the absorption of more solar
radiation by impermeable surfaces, the rise of temperatures in urban areas and the exacerbation of
the urban heat island effect, which has had a variety of negative impacts on human societies. Blue green space has the “cold island effect”, which is opposite to the heat island effect, and can play an
ecological role in regulating urban microclimate and enhancing urban biodiversity. However, there is
a lack of in-depth quantitative analysis and research on the cold island effect of waterfront blue-green
spaces in large cities in China.
In this study, it takes Wuhan, Nanjing and Hangzhou as cases of three cities with developed
water systems, and analyzes the cold island effect of blue-green space in three cities with developed
water systems (Wuhan, Nanjing and Hangzhou) with the help of multi-source remote sensing and
GIS spatial data, and analyzes the cold island effect of blue-green space in the three cities (Wuhan,
Nanjing and Hangzhou) with the help of a machine-learning regression decision-tree model to
reveal the relationship between blue-green space factors and heat island effect in different waterfront
spaces. Factors interact with the heat island effect. In this study, the typical urban center with rich
water system is taken as the study area of the cold island effect, and the neighborhood unit is taken
as the smallest unit, which highlights the concentrated investigation of the cold island effect of the
blue-green space along the waterfront of the city, meanwhile, the decision tree regression is used to
solve the problem of the large scope of the study, the number of data is very large, and there is the
problem of the linear influence between the respective variables. The results show that the strong
cold island effect areas in the waterfront blue-green space in Wuhan are located around large lakes
such as Donghu Lake and Tangxun Lake, while the south bank of the Qinhuai River is in Nanjing,
and the south side of the middle part of the Qiantang River and the south side of the West Lake are
in Hangzhou. Decision tree regression results show that there are also differences in the landscape
morphology factors that play a dominant role in different cities, and that the watershed morphology
of different cities plays an obvious role in influencing the urban heat island. In Hangzhou and
Nanjing, water surface rate and green space rate are the dominant factors, and Hangzhou is jointly
affected by them in 72.9%, while Nanjing is 61.8%. In Wuhan, water surface rate and water shape
index are the dominant factors, which are jointly influenced by 63.5%.
In view of the different waterfront blue-green spatial characteristics of the three large cities with
abundant water systems, different cities should put forward targeted optimization strategies. Wuhan
can continue to improve the connectivity of large water bodies to further form an overall blue-greenspace system. Hangzhou should appropriately increase the number of artificial lakes and connect scattered water bodies in the city, and further enrich the
shape of waters and improve the overall network of blue-green space in the process of development and construction of new districts and improvement of the
water network. Nanjing should continue to give full play to the existing advantages of green space, improve the diversity of green space form boundaries,
and at the same time need to set up a combination of waterfront green space system, break the continuous urban heat island, and improve the overall cooling
capacity. With the development of the economy, in recent years, urban development has become more concerned about ecological livability, harmonious
coexistence of man and nature. For big cities that have been built, it is difficult to adjust the blue-green space pattern by blindly altering and rebuilding on
a large scale, but it is possible to start from the more microscopic waterfront blue-green space and carry out small-scale renovation, so as to achieve the
ecological planning of blue-green space without large-scale demolition and alteration, and to achieve the purpose of mitigating the urban heat island effect.
Based on multi-source remote sensing data, GEE, machine learning-based decision tree regression has a greater potential in the analysis of cold island
effect in blue-green space of different water network cities, and the high-precision spatial dataset generated by the study enriches the refined quantitative
analysis of the cold island effect of blue-green space of the city, and at the same time provides scientific references for the planning and regulation of blue green space of the waterfront and the management and control of the thermal health risk of other similar cities. The results of this study can provide useful
references for the scientific planning and design of waterfront blue-green spaces in large cities in China. |
| Key words: Urban Cold Island Intensity Blue-Green Space Urban Water Network Shape Decision Tree Regression |
