Michigan Technological University
Synergistic time-series analysis of satellite data: applications to volcano monitoring
Abstract
dc:description.abstractSatellite remote sensing measurements represent a global data source for the investigation of a variety of geophysical phenomena. This research exploits over 10 years of data available from the Ozone Monitoring Instrument (OMI) and two Moderate Resolution Imaging Spectroradiometer (MODIS) sensors, maintained by NASA, to investigate ongoing volcanic eruptions. Time-series analysis techniques are increasingly being employed in the assessment of volcanic emissions in satellite data, and investigation of these data indicated a cyclical characteristic common to these instruments (period of ~2.3 days) resulting from variations in viewing angle generated by sun synchronous satellites maintained in Low Earth Orbit (LEO). A secondary feature was also identified (periods of 3.2- and 7.9-days) in data affected by the OMI Row Anomaly (ORA). Following the development and implementation of a correction methodology, the feasibility of monitoring volcanic eruption dynamics by satellite instruments was conducted. The initial assessment of Soufriere Hills Volcano (SHV; Montserrat) indicated the capabilities of the incorporated methodology through the identification of correlating cycles (8-, 12-, 54- and 172-days) in this work when compared to ground-based analyses employed over a similar time frame. The subsequent extension of this work to other sites of volcanic interest resulted in the identification of dominant cyclicity, external forcing, variations in cyclical dynamics over time and the classification of major SO2 emissions source in a complex volcanic region. These results indicate the flexibility of the developed methodology in the assessment of volcanic features over a variety of timescales and activity types. In addition to time-series analysis satellite remote sensing has potential applications to near real time assessment of volcanic systems. To exploit the operational nature of the OMI sensor an automated detection algorithm was developed, facilitating the identification of SO2 plumes with an accuracy of over 80%. Through refinement of the incorporated model when a detection threshold of 400 tons was applied to plume detection, the accuracy of the model increases to ~97% of correctly classified events. The research conducted here demonstrates the potential of moderate resolution satellite-based remote sensing data for the identification and assessment of ongoing dynamic characteristics and near real-time monitoring of volcanic systems.
Degree
thesis:*- Name thesis:degree_name
- Doctor of Philosophy in Geology (PhD)
- Level thesis:degree_level
- Campus Access Dissertation
- Discipline thesis:degree_discipline
- Department of Geological and Mining Engineering and Sciences
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Flower, Verity
- Committee members dc:contributor.committeemember
-
- Louisa J. Kramer
- Gregory P. Waite
- Robert Wright
- Contributors dc:contributor
-
- Simon A. Carn
Subjects
dc:subject × 10Identifiers
dc:identifier.*- Identifier
- https://digitalcommons.mtu.edu/etdr/16