Precision and accuracy of Earth Observing-1 Advanced Land Imager data for semiarid vegetation studies

Fast and effective land monitoring is particularly important in and and semiarid lands where changes in water resources alter the landscape on a large scale. In previous work, we have used Landsat TM data and Spectral Mixture Analysis to successfully estimate vegetation live cover to +/-4.0% live cover (%LC). In this paper we compare estimates of %LC derived from Earth Observing-1 (EO-1) Advanced Land Imager (ALI) data to estimates derived from field based analyses. Results show that ALI estimates of %LC had similar precision to estimates derived from ETM+ data; these values were +/-5.61 %LC for ETM+ and +/-6.15 and +/-6.44 %LC for ALI using all 9 bands and only 6 bands, respectively. The only significant difference sensor performance, was the quality of multispectral band registration. ALI was found to be internally inconsistent in that each of the four SCA image swaths were co-registered to varying precision. This feature is most significant in regions of high vegetation variability and would also be problematic for studies of annual vegetation change.