Low spatial resolution thermal monitoring of volcanoes from space

Harris, Andrew John Lang (1996). Low spatial resolution thermal monitoring of volcanoes from space. PhD thesis The Open University.

DOI: https://doi.org/10.21954/ou.ro.0000f7b7

Abstract

Low spatial resolution thermal data from the Advanced Very High Resolution Radiometer (AVHRR) have been available for any cloud-free sub-aerial volcano at least 4 times a day since 1979. Although extensively used to monitor ash plumes injected into the atmosphere by volcanic eruptions, this cheap, directly and regularly available data source has never been thoroughly examined to determine its utility for real-time monitoring of high temperature volcanic activity occurring at the surface. This thesis is therefore concerned with developing techniques to locate, measure and monitor high temperature volcanic phenomena in AVHRR data.

To locate high temperature phenomena, I develop an automated thresholding algorithm which roams an image flagging thermal anomalies. The major advantage of this technique over existing thresholding techniques developed for fires is that it requires no manual input. By running the algorithm on 90 images acquired during the 1991 to 1993 effusive eruption at Mount Etna, pixels containing active lava or high temperature open vents were located, with the program completing within 1 second. This permitted lava flow location to within ±1 km of known topographic features, roads and towns. The flexibility of the thresholding algorithm to detect other high temperature sources, over a variety of regions and seasons, was demonstrated by applying the algorithm to images containing wild-fires burning along the east coast of Australia during January 1994.

For quantitative analysis, I develop a new approach capable of estimating the size and temperature of at least three sub-pixel thermal components. This approach enables lava flow area and thermal flux to be constrained more accurately than is possible using established techniques which allow measurement of just two thermal components.

Using the area and thermal flux results, I refine an algorithm to estimate lava effusion rates. This in turn allows calculation of lava flow field cumulative and total volumes. Application to data for effusive activity at Etna during 1985 and between 1991 and 1993 gives mean effusion rates of <8 m³ s^-1 during both eruptions, building flow fields of 16-29 x 10⁶ and 220-300 x 10⁶ m³ respectively. Both volume estimates are in good agreement with ground-based estimates of 19 x 10⁶ and 235 x 10⁶ m³ respectively. In both cases, time series of area, thermal flux, effusion rate and cumulative volume data reveal phases of waxing and waning eruptive activity, in close agreement with those observed from the ground.

I also present simple approaches which use spectral radiance from the volcanic source to monitor volcanic activity through time and to distinguish between different types of volcanic phenomena. At Stromboli, troughs and peaks in a spectral radiance time-series respectively correlate with periods "normal" explosive activity and less frequent effusive events. A scatter plot of spectral radiances from the volcanic source acquired in two thermal wavebands, one in the mid-infrared and one in the thermal infrared, allows lava lakes, flows and open vents to be distinguished according to their position in the scatter.

I use a number of case studies, drawn from effusive activity at Etna, Krafla, Togo, Cerro Negro and Erebus, as well as pyroclastic and degassing activity at Lascar, Vulcano, Etna and Stromboli, to show how AVHRR can be used to measure and monitor high temperature volcanic phenomena. If data analysis were carried out on reception, regular location and monitoring of on-going effusive eruptions, as well as persistent high temperature phenomena, would be possible.