Download files
Abstract
The urban heat island (UHI) is one of the most documented manifestations of urbanisation and the subject of intensive research over recent decades. Australia, as many other highly urbanised countries, has developed policies and strategies to promote more compact settlements; however, new urban development is characterised by higher urban densities and larger proportion of impervious surfaces that can potentially intensify UHIs. Among mitigation strategies, green infrastructure (GI) has proved effective in reducing urban temperatures. However, more research is needed to determine which compositions, amounts and spatial distributions are more effective in providing cooling benefits. This PhD research responds to this need by proposing a new taxonomy of green infrastructure typologies (GITs) to classify urban landscapes into 34 standard classes. Very high-resolution thermal, spectral imagery and LiDAR data were employed to examine the relationships between functional, structural and configurational descriptors of GI and the diurnal and nocturnal thermal patterns across the Sydney metropolitan area. Remote sensing data were collected by aircraft in February 2013 (summer) and August 2012 (winter) in calm, clear and dry conditions. This study demonstrates the applicability of the proposed methodological framework and classification scheme in urban climatology by analysing the inter- and intra-variability of land surface temperatures (LSTs) among typologies. Results show that water bodies, well irrigated grasses and aligned and clustered trees provide the largest cooling effects at daytime. However, at night well irrigated grasses are much cooler than forested areas, while water features are among the warmest GITs. It was also found that: (1) the composition and abundance of land covers is more influential in LSTs than the spatial distribution; (2) the lack of irrigation significantly affects the cooling capacity of vegetation, and (3) the cooling effect of vegetation is significantly outweighed by the warming effect of surrounding impervious surfaces. Several statistical models were produced for an accurate prediction of LSTs based on the individual contributions of derived GI parameters. Resulting estimates were employed to propose key principles, mitigation strategies and guidelines that can be implemented by researchers, governments and practitioners to prioritise greening interventions, improve urban microclimates and mitigate the urban heat more effectively.