Application of satellite hyperspectral remote sensing for achieving sustainable development in Local gGovernment Authority areas of NSW, Australia

The Local Government Act 1993 is a key statute which governs the existence and operation of Local Government Authorities (LGAs) of NSW. The act mandates the incorporation of quadruple bottom-line decision making by LGAs including the protection and enhancement of natural lands within their jurisdictions. Natural lands contain ecosystems which include vegetation, which are important for maintaining a healthy urban environment. Hyperspectral Remote Sensing (HRS) images can be utilised to identify unique spectral signatures of vegetation types at given locations. The development of satellites for commercial Hyperspectral image acquisition provides new opportunities to achieve cost effective regular monitoring and mapping of land cover types in comparison to imagery acquired through airborne methods. The spatial resolution of Satellite HRS imagery generally vary between 30-60m. Spectral Mixture Analysis (SMA) methods can be used to identify land cover fractions including vegetation that occur at subpixel level. Hence, satellite HRS images and SMA analysis provides environmental practitioners new, cost effective opportunities to map vegetation species at subpixel level, more frequently, at higher resolutions; which are important for achieving higher levels of environmental sustainability in urban environments. This study using Hyperspectral imagery was completed in 2012 in The Nurrangingy Reserve situated in Doonside NSW. The image used was acquired in 2001 by the Hyperion sensor on board NASA’s EO1 satellite. The Sequential Maximum Angle Convex Cone (SMACC) SMA method was used to analyse subpixel vegetation fractions. The field verification of the image analysis results via 16 Endmember sites (Endmember site is an area where the corresponding image pixel represents a single species) and 9 field verification sites resulted in 85% overall unmixing reliability with a 78% Endmember identification rate and a 68% Species identification rate. Despite the age of the dataset used in the study, the results depicted considerable promise for the workflow employed for identification of subpixel vegetation fractions. This paper presents the methodology of the study, the findings and their potential improvements and discusses the relevance of the outcomes in relation to conservation of urban forest ecosystems and new opportunities that arise for enhancing sustainable development in Local Government Authority areas through the incorporation of environmental investment.