Spatial structure of surface urban heat island and its relationship with vegetation and built-up areas in Melbourne, Australia

Yashar Jamei; Priyadarsini Rajagopalan; Qian Chayn Sun

doi:10.1016/j.scitotenv.2018.12.308

Spatial structure of surface urban heat island and its relationship with vegetation and built-up areas in Melbourne, Australia

Sci Total Environ. 2019 Apr 1:659:1335-1351. doi: 10.1016/j.scitotenv.2018.12.308. Epub 2018 Dec 29.

Authors

Yashar Jamei¹, Priyadarsini Rajagopalan², Qian Chayn Sun³

Affiliations

¹ School of Property, Construction and Project Management, RMIT University, Melbourne, Australia. Electronic address: Yashar.jamei@rmit.edu.au.
² School of Property, Construction and Project Management, RMIT University, Melbourne, Australia.
³ School of Science Cluster, Department of Geospatial Science, RMIT University, Melbourne, Australia.

PMID: 31096344
DOI: 10.1016/j.scitotenv.2018.12.308

Abstract

Due to the intensity of urban development around the world, there is an increasing body of studies attempting to investigate urban heat island (UHI) in various spatial and temporal scales. In surface heat urban island (SUHI) studies, extended periods of time, broader regions and local government area (LGA) level have become more crucial and will shed light on causes of UHI. Moreover, the spatial pattern and structure of SUHI will be useful for policy-makers to develop mitigation strategies. This study focused on three objectives. Firstly, analyzing land surface temperature (LST), normalized difference built-up (NDBI) and vegetation (NDVI) indices. Secondly, investigating interrelationships among LST, NDVI, and NDBI. Thirdly, identifying LST patterns in the Melbourne metropolitan area. These objectives were achieved through three different methods. The modified automatic mapping method for the first objective, the correlation analysis for the second, and spatial statistical methods for the third. The methodological innovations of this study were considering LGA in interrelationship analysis among LST, NDBI and NDVI, and calculation of NDVI for each acquisition date. The results indicated that the clustering pattern of LST expanded toward the north-west and south-east during the period of the study. Furthermore, the north-west part of the city has the highest positive (0.6) correlation between NDBI and LST, and the south-east part of the city has the lowest negative (-0.8) correlation between NDVI and LST. The most significant increase and decrease in mean LST happened respectively from January 6th to 22nd 2017, and January 14th to 30th January 2014. The temperature degree altered from 19.61 °C to 27.86 °C in inner western suburbs, and from 35.49 °C to 26.88 °C in most LGA's. These findings are critical for planners to localize UHI mitigation action plans, target hot spots in LGA's and allocate resources to respond to the adverse effect of UHI.

Keywords: Land surface temperature (LST); Melbourne metropolitan area; Spatial distribution; Surface urban heat island (SUHI).