Estimation of Land Surface Temperature and Urban Heat Island effect for Australian urban centres

Environmental Management

Devereux, Drew ; Caccetta, Peter

CSIRO Data61

DOI  :  10.4225/08/5a0497a855f6f
RP-ID  :  EP173542

学科分类：地球科学(综合)

澳大利亚|英语

来源: CSIRO Research Publications Repository

【 摘 要 】

This report describes the generation of land surface temperature (LST) and urban heat island (UHI) estimates for major Australian urban centres.Landsat 8 thermal infrared imagery was obtained for each urban centre for as many overpasses as possible during the summer of 2015/16. Each image was processed to LST using the single channel method (Jiménez-MuñozSobrino, 2003, Jiménez-Muñoz et al, 2009). The required atmospheric parameters were obtained from publicly available observations by the Australian Government Bureau of Meteorology (BOM). The required land surface emissivity (LSE) values were estimated using the NDVI approach (SobrinoRaissouni 2000). Images were then averaged to obtain an estimate of typical summer LST.The UHI is a measure of the deviation of urban temperatures relative to a non-urban baseline. Native vegetated sites were used to establish the baseline. This was achieved by estimating a first-order fit to the temperature of native vegetation within and around each urban centre. This fit captures any broad-scale temperature trend that is likely independent of urbanisation, such as cooling with increased latitude or proximity to the coast. After subtracting this fit, the residuals may be interpreted as showing finer scale deviations from this trend, including deviations attributable to urbanisation of the landscape.The quality of results vary with the availability of cloud- and haze-free data. Due to the limited number of satellite overpasses for each region for the time-period considered, cloud- and haze-free data is not guaranteed for each region, in which case the best available data was used. In locations where cloud-affected imagery was used, artefacts in the results may remain due to limitations in the accuracy of the cloud masks. UHI estimates for the cities including Sydney, Perth, Melbourne and Adelaide were relatively free of cloud induced artefacts, whereas Brisbane, Hobart and Cairns contain some artefacts which limit their utility.Where adequate cloud-free data was available, the results may be safely interpreted as showing significant temperature increases across all urban areas, with the largest increases occurring in areas with extensive hardpan, such as airports and industrial areas. Although the data is too coarse to allow examination of fine scale features such as individual houses and roads, there is some evidence of temperature increases along major roads and above some large industrial or commercial buildings such as shopping centres and distribution warehouses.