Discriminating Compositional Variations on the Silicic Domes of Medicine Lake Volcano, CA, with the New Airborne Hyperspectral MODIS/ASTER Simulator
1
1 Department of Geology, Arizona State University, Tempe, AZ, 85287, USA
2 Department of Geology and Planetary Science, University of Pittsburgh,
Pittsburgh, PA, 15260, USA
Subtle variations in the surface composition of active lava domes and flows can signal a change in the emplacement conditions and potential explosivity of a volcano. Observing and monitoring these variations may be useful for volcanic hazard mitigation, in addition to understanding eruption dynamics and evolution. Direct sampling and observation, though highly desirable, do not provide good spatial and temporal coverage. Further, these avoids these problems by utilizing sensors that are well removed from volcanic hazards.
The MASTER (MODIS/ASTER Airborne Simulator) instrument is a new NASA hyperspectral sensor that acquired data over the silicic lava domes at Medicine Lake Volcano, CA, in September, 1999. Data were simultaneously recorded in 50 spectral channels, from the visible through thermal infrared (TIR), at a spatial resolution of approximately 5 meters. The high spectral and spatial resolutions of these data allow us to observe small compositional variations in the SiO2 content. With ten bands in the TIR (8 –12 mm), the data also provide better constraints on the spectral masking effects of vitrification and vesiculation associated with glassy lavas. The older Thermal Infrared Multispectral Scanner (TIMS) sensor had only six spectral bands in the TIR, which is insufficient to separate these effects. The MASTER instrument also gives us a first look at silicic lavas in the mid-wave infrared (3 – 5 mm).
For this study we focused on Big Glass Mountain, which is compositionally zoned from dacite (63.9 wt % SiO2) to rhyolite (74.1 wt % SiO2), and in some areas has complex mixing between these units. Initial image processing using a spectral deconvolution algorithm with four chemical end-members shows silica variability most distinctly in the northeastern dacite lobe of Big Glass Mountain. Principal component transformations and spectral ratioing reveal convoluted spatial patterns that are potentially linked to mixing during flow advance. Previous field mapping and TIMS image analysis allows an excellent comparison with MASTER derived images. Future work will include direct sampling, mapping, and chemical analysis, as well as a comparative analysis of data from the recently launched Advanced Spaceborne Thermal Emission and Reflectance Radiometer (ASTER). This instrument will provide good temporal coverage at lower spectral and spatial resolutions, and ultimately allow monitoring of active silicic domes and lava flow hazards. We also plan to use laboratory simulations to model the physical mixing processes responsible for the observed distribution of lava compositions.
------------------
Submitted: IAVCEI General Assembly, Bali, Indonesia
Date: July, 2000