Discrimination of arid vegetation composition with high resolution CASI imagery.

Abstract

CASI (Compact Airborne Spectrographic Imager) airborne imagery, with high spectral and spatial resolution, was evaluated for the discrimination of composition and variation in arid vegetation at Fowlers Gap Arid Zone Research Station in western New South Wales. The imagery was calibrated to surface reflectance using field reference spectra collected near the time of the overflight, and analysed in relation to reflectance spectra of plants at Fowlers Gap. Maps showing abundance of total perennial vegetation, chenopod shrubs and trees, were produced using methods that separated the vegetation contribution from mixed-pixel responses. Results of these analyses were compared with field data on percentage ground cover for 85 one hectare sample plots, collected within four weeks of the overflight. In all cases, the cover of vegetation that was discriminated and mapped was less than 25%. The study demonstrates that high-spectral resolution imagery, combined with new approaches to image analysis, offers considerable scope for discrimination of vegetation variation in arid landscapes. It is possible to spectrally discriminate and map the abundance of several functional vegetation components, even in sparse vegetation, and this information is particularly relevant for management applications. As imagery from high-spectral resolution sensors, both air- and satellite-borne, becomes more readily available in Australia, the benefits to be derived from these data will be more widely applied. Key words: arid vegetation, species composition, hyperspectral imagery, remote sensing, vegetation spectra