Utilizing world urban database and access portal tools (WUDAPT) and machine learning to facilitate spatial estimation of heatwave patterns

Shi, Yuan ORCID: 0000-0003-4011-8735, Ren, Chao, Luo, Ming, Ching, Jason, Li, Xinwei, Bilal, Muhammad, Fang, Xiaoyi and Ren, Zhihua (2021) Utilizing world urban database and access portal tools (WUDAPT) and machine learning to facilitate spatial estimation of heatwave patterns. URBAN CLIMATE, 36. p. 100797.

Text
UCLIM-2021-Utilizing world urban database and access portal tools (WUDAPT) and machine learning.pdf - Author Accepted Manuscript
Download (2MB) | Preview

Official URL: http://dx.doi.org/10.1016/j.uclim.2021.100797

Abstract

Climate change lead to more intense, higher frequent and prolonged heat extremes. Understanding the spatial pattern of heatwave is vital for providing the corresponding weather services, making climate change adaptation strategies and heat-health actions. In this study, we present an approach to estimate the heatwave spatial patterns by utilizing the WUDAPT Level 0 data and machine learning. The analysis is based on two years (2009 and 2016) of air temperature data from 86 meteorological monitoring stations in Guangdong province of China, a subtropical region with frequent hot and sultry weather in summer. First, heatwave conditions were quantified by calculating the number of hot days and frequency of heatwave events in each year and used as the response variables. Then, random forest models were built by using a geospatial dataset consisting of WUDAPT and urban canopy parameters (UCP) as predictor variables. Based on the resultant models, spatial patterns of heatwave were estimated and mapped at 100 m spatial-resolution. The results show that this approach is able to estimate heatwave spatial patterns using open data and inform urban policy and decision-making. The study is also a new perspective and a feasible pathway of utilizing WUDPAT Level 0 product to facilitate urban environment applications.

Item Type:	Article
Uncontrolled Keywords:	Heatwave, Random forest, WUDAPT, Machine learning, Spatial estimation
Divisions:	Faculty of Science and Engineering > School of Environmental Sciences
Depositing User:	Symplectic Admin
Date Deposited:	25 Jul 2022 14:37
Last Modified:	18 Jan 2023 20:55
DOI:	10.1016/j.uclim.2021.100797
Related URLs:	Author Publisher
URI:	https://livrepository.liverpool.ac.uk/id/eprint/3159338