Determining subpixel thermal anomalies and heat flux of Cerro Negro Volcano, Nicaragua: A TIR deconvolution approach

S. Rose, Department of Geology and Planetary Science, University of Pittsburgh, Pittsburgh, PA

M. Ramsey, Department of Geology and Planetary Science, University of Pittsburgh, Pittsburgh, PA

Deconvolution, or the process of extracting information below inherent spatial or spectral resolution of a remote sensing dataset, can be an invaluable tool for understanding many volcanological processes and hazards. For example, this technique can be used to understand volcanic unrest and associated variations in processes such as thermal flux, surface chemistry, and lava texture. This study utilizes satellite and ground-based thermal infrared (TIR) remote sensing to identify and monitor the thermal flux on and around Cerro Negro Volcano. Cerro Negro lends itself well to a remote sensing study due to its exposure, relative safety for ground-based data collection, and the variation of surface alteration and temperature. The volcano is located within the El Hoyo-Las Pilas volcanic complex of northwestern Nicaragua, ~21 km east of the city of Leon. Over the past 150 years, this young stratovolcano has been intermittently active, with the most recent major eruption occurring in 1995... Multispectral spaceborne TIR data from the ASTER instrument has been separated into surface temperature and emissivity, with the latter used to map petrologic variations on the eruptive products. However, the accuracy of this mapping depends greatly on the deconvolution model as well as knowledge of the subpixel temperature structure of each pixel. A thermally-heterogeneous pixel can produce significant errors in the derived emissivity. We document a new technique that solves this complication using both modeling and field-based data collection. Fieldwork planned in early 2006 will utilize a FLIR camera to validate the satellite data as well as measure the volumetric flux of heat, water, and gas species from the fumaroles. Upon completion, we expect to produce a TIR deconvolution algorithm capable of extracting accurate surface temperature and emissivity on a subpixel scale. Such data can then be used to model gas heat flux from a volcanic vent or fumarole field.

------------------
Submitted: Cities on Volcanoes 3 Meeting (2006)
Date: January 23 - 27, 2006