Satellite thermal remote sensing of active volcanoes : understanding volcanic processes and application to automated monitoring algorithms

• Main Author: Manen, Saskia Marjoleine van
• Format: Thesis
• Language: unknown
• Published: 2009
• Subjects: Kamchatka; Alaska
• Online Access: https://oro.open.ac.uk/65638/; https://oro.open.ac.uk/65638/1/518172.pdf; https://doi.org/10.21954/ou.ro.00010066

Quantitative retrospective analysis of coarse resolution thermal time series and multiparametric studies are used to improve the understanding of volcanic activity and to advance existing monitoring algorithms. Four months of data from the 2006 Augustine (Alaska) eruption show that thermal activity correlates with reduced displacement, enabling qualitative interpretation of the surface activity. Differences in 5 years of thermal time series from Bezymianny and Kliuchevskoi (Kamchatka) relate to the contrasting physical properties of Bezymianny’s andesitic volcanism and Kliuchevskoi’s basaltic lavas and their 2 km difference in summit elevation, which causes systematic differences in the ambient temperature into which volcanic material is erupted. Using 7 years of data at Oldoinyo Lengai (Tanzania) potential external forcing of magmatic activity by solid Earth tides was examined, but no correlation was found. Time series of thermal satellite data spanning fifteen years were examined for Bezymianny, Shiveluch, Kliuchevskoi and Karymsky. Bezymianny and Shiveluch (Kamchatka) showed varying complex oscillatory behaviours associated with a systematic process leading to explosive conditions. At Bezymianny, explosions were largely preceded by increases in radiant temperature. Karymsky displayed a strong thermal signal from an extruding lava flow, but less so from its frequent Strombolian behaviour. Kliuchevskoi exhibited a limited range of activity patterns owing to less pressurization of the magmatic system compared to its andesitic counterparts. Using these results two successful algorithms were developed: (1) distinguishing Bezymianny and Kliuchevskoi within a single image and (2) forecasting explosions at Bezymianny. Both these algorithms retrospectively have an 89% success rate. This could be improved upon by analysis of higher spatial resolution or longer datasets. Detailed analysis of long term thermal time series is essential as much valuable information regarding precursory activity, eruption styles and how ...