How far can imaging spectroscopy go in identifying the cause of vegetation stress around landfill sites?

Recent improvements in the spatial and spectral resolutions of remote sensing imaging instruments have increased vastly their ability to detect subtle variations in the health of vegetation and in the soil in which it is growing. This capability, coupled with the unique ability of remote sensing data to produce synoptic maps of a land surface, make them a potentially powerful tool for monitoring the spread of contaminants within a soil-body. The aims of this paper are to review the potential for this application of imaging spectroscopy, and to describe current work which is aimed at assessing how well spectrometric data can distinguish the cause of vegetation stress. This work analyses the residual spectrum which results from subtraction of a healthy vegetation spectrum from that observed and using controlled field trials, deduces characteristic spectral elements which are indicative of different causes of stress. These are then fed into an artificial neural network, which interrogates each pixel in a spectrometric image and returns an assessment of the health of the vegetation therein. The application to landfill sites is central to this work, but future work expanding the application of such a method to derelict brownfield sites is discussed.