Integration of thermal infrared satellite data with ground-based geophysical data for understanding volcanic processes

This dissertation examines the tools and applications available for investigating volcanic phenomenon using satellite thermal infrared remote sensing data. Various algorithms have been developed to automatically detect volcanic thermal anomalies. A contextual (VAST), fixed threshold (MODVOLC) and temporal (RAT), the three main types of algorithms, are compared to determine how effective they are at detecting thermal anomalies caused by various types of volcanic activity (lava flows, lava domes, strombolian activity, and fumarolic activity). Each of the algorithms operates with the highest accuracy for the types of activity that they were designed to detect and no algorithm is 100 percent accurate. With the current data restrictions no algorithm ever will be, therefore user interaction is key.Results from the automated algorithms are then applied to determine discharge rates and cumulative volumes of erupted lava during the Stromboli 2007 eruption. Blinding applying the result can cause errors up to an order of magnitude with the main cause of errors coming from the inclusion of cloudy data and not identifying the most radiant pixels. From the manual results it was determined the 2007 eruption was a typical Strombolian effusive eruption caused by tapping a pressurized magma source.The satellite derived discharge rates and SO2 flux supply rates are then calculated and compared at Etna from 2002-2006. Differences in the supply rate of magma and erupted volume of lava occur from eruption to eruption and also vary throughout individual eruptions, indicating a complex supply system within Etna.Thermal satellite data is also used to estimate plume heights at Tungurahua volcano from 2006-2008. Heights are compared with acoustic power. Good correlation between plume height and acoustic power was found. By integrating the two data sets it is possible to distinguish between different eruption styles and aids in classification of eruption types.Although satellite data is a useful tool for volcano monitoring image analysts are needed for the most accurate quantitative results. Incorporating satellite data with other geophysical measurements allow insights into the eruption processes that would not otherwise be possible. Therefore it is important that every possible resource is used in volcanological studies.