The primary goal of this research was to develop an innovative survey method that could be applied to the prediction, confirmation, and understanding of archaeological sites utilizing aerial and ground‐based remote sensing methods in a combined fashion. By drawing on the resources of the University of Colorado Department of Anthropology, the Center for the Study of Earth from Space (CSES), and the Institute of Arctic and Alpine Research (INSTAAR), this combined remote sensing method has been used successfully in locating and evaluating ancient chert quarries in the Horse Prairie region of southwestern Montana. The quarries in this region have served as a source of quality raw material for over 10,000 years, and are found along veins of hydrothermal alteration and in deposits of till.
By combining ground‐truth data based on electromagnetic induction surveying and conventional aerial photography and principal components image analysis of Landsat imagery, various spectral class signatures were created and then reapplied to the raw spectral data in a supervised maximum‐likelihood multispectral classification. Subsequent field checking and spectroscopy data illustrate that this classification successfully predicted the location of previously undocumented quarry areas and also suggests that the quarries are highly restricted geographically within the region. The data obtained from this research will also contribute understanding to such problems as the nature of lithic resource procurement and associated economic relationships between the Great Plains and Great Basin areas of the western USA. In addition, this research method will be applicable to a variety of alpine, Arctic, and semi‐arid regions where ground surfaces are exposed and spectral signatures can be established. © 1996 John Wiley & Sons, Ltd.