背景(考古学)
环境科学
城市化
城市热岛
贝叶斯多元线性回归
灵敏度(控制系统)
线性回归
多元统计
强度(物理)
地理
自然地理学
气象学
统计
数学
量子力学
经济增长
物理
工程类
经济
考古
电子工程
作者
Yunfei Li,Bin Zhou,Manon Glockmann,Jürgen P. Kropp,Diego Rybski
标识
DOI:10.1016/j.scs.2021.103146
摘要
In this study we analysed the multi-annual (2002–2011) average summer surface urban heat island (SUHI) intensity of the 5000 largest urban clusters in Europe. We investigated its relationship with a proposed Gravitational Urban Morphology (GUM) index that can capture the local context sensitivity of SUHI. The GUM index was found to be an effective predictor of SUHI intensity. Together with other urban factors we built different multivariate linear regression models and a climate space based geographically weighted regression (GWR) model that can better predict SUHI intensity. As the GWR model captures the variation of influence from different urban factors on SUHI, it considerably outperformed linear models in predicting SUHI intensity in terms of R2 and other statistical criteria. By investigating the variation of GWR coefficients against background climate factors, we further built a nonlinear regression model that takes into account the sensitivity of SUHI to regional climate context. The nonlinear model showed comparable performance to that of the GWR model and it prevailed against all the linear models. Our work underlines the potential of SUHI reduction through optimising urban morphology, as well as the importance of integrating future urbanisation and climate change into the implementation of urban heat mitigation strategies.
科研通智能强力驱动
Strongly Powered by AbleSci AI