RAMSEY, Michael S., and FINK, Jonathan H., Department of Geology, Arizona State University, Box 871404, Tempe, AZ 85287-1404, ramsey@elwood.la.asu.edu and jonathan.fink@asu.edu

Remote sensing has provided multispectral spatial and temporal coverage for both geologic mapping and monitoring at hundreds of volcanic centers throughout the world. With the launch of the ASTER instrument in 1998 aboard the first Earth Observing satellite, world-wide multispectral data from the visible to the thermal infrared will become available for the first time. In the past, volcanologists have used remote sensing to measure characteristic parameters of lava flows and domes such as length, aspect ratio, and roughness, all of which being indicators of the more fundamental conditions present. One parameter, lava vesicularity, is of interest since it provides information regarding the volatile content, emplacement time, and internal structure. Regardless of the wavelength used however, remote sensing has not been the primary tool to derive surface vesicularity due to factors such as bubble size and the scattering physics.

Pumiceous textures do produce significant variations in thermal emission spectra that are clearly distinguishable. For example, the distinct spectral feature of volcanic glass is commonly muted due to the overprinting of the vesicles acting as featureless blackbody radiators. Young silicic domes, having surface textures that vary from dense glass to highly vesicular pumice, provide an excellent test for techniques designed to extract the degree of vesiculation. Spectrally, these textures describe a continuum consisting of two pure endmembers, glass and bubbles. Assuming that the emission from both combines linearly in direct proportion to their areal percentage, the spectra can then be deconvolved to estimate the abundances of each endmember.

A linear deconvolution model was developed and applied to airborne, thermal infrared data of several Holocene silicic domes and basaltic flows in central and northern California. With the glass and a blackbody as spectral endmembers, the retrieval algorithm was utilized to produce a vesicle-percentage image for the domes. Average values extracted from regions across the surface compared well with the previously reported values and further thin section analysis. The applicability of such a technique for basaltic compositions was also explored. The use of a rapid analysis technique such as this for future thermal infrared data will provide insights into the degassing state of an active flow and aid in determining the potential of subsequent eruptive activity.

Key words: volcanology, vesicularity, thermal infrared, remote sensing, spectral unmixing

------------------
Presented at: Geological Society of America Cordilleran Meeting
Date: 1997