Evaluation of Alternative Sensors for a “landsat-based” Monitoring Program

In Australia, Landsat imagery is currently used in a number of regional and national monitoring projects to provide maps of the extent and change in area of perennial vegetation. They provide basic information for conservation, land management and for modelling carbon flux and water use. With the looming gap in Landsat data continuity it is timely to consider the issues involved in using data from other sensors to continue these monitoring programs. In the context of the Australian Greenhouse Office Land Cover Change Program, this paper describes the issues and quantifies the effects of using Spot 4 and Landsat 7 SLC-OFF images instead of the current Landsat 5 images. The data ‘pre-processing’ issues investigated include ortho-rectification, calibration and terrain illumination correction. Overlapping sets of images from three different geographic regions were processed to assess logistical and technical issues. The ability to discriminate between classes of interest is considered in the context of the forest (>20% canopy cover) monitoring. The accuracy of the change products from ‘mixed’ sensor time series analysis is also discussed. Both the accuracy of the products from each step in the processing and the cost in terms of processing time and complexity are reviewed.