Publications

Bonny, E; Wright, R (2017). Predicting the end of lava flow-forming eruptions from space. BULLETIN OF VOLCANOLOGY, 79(7), 52.

Abstract
Although the volcanological community places great emphasis on forecasting the onset of volcanic eruptions, knowing when an effusive eruption will end is just as important in terms of mitigating hazards. Wadge (J Volcanol Geotherm Res 11: 139-168, 1981) postulated that the onset of an episodic, lava flow-forming basaltic eruption is characterized by a rapid increase in effusion rate to a maximum, before decaying over a longer period of time until the eruption ends. We used thermal infrared remote-sensing data acquired byNASA'sMODerate Resolution Imaging Spectroradiometer (MODIS) to derive time-averaged discharge rate (TADR) time series using the method of Harris et al. (J Geophys Res 102(B4): 7985-8003, 1997) for 104 eruptions at 34 volcanoes over the last 15 years. We found that 32 eruptions followed the pattern described by Wadge (J Volcanol Geotherm Res 11: 139-168, 1981). Based on the MODIS-derived maximum lava discharge rate and a decay constant that best fits the exponential waning phase (updated as each new MODIS TADR observation is added to the time series), the time at which the discharge equals zero, and thus the point at which effusion ends, can be predicted. The accuracy of the prediction improves with the number of data points so that, in the ideal case, the end of effusion can be retro-casted before half of the eruption duration has passed. This work demonstrates the possibility of predicting when an eruption will end using satellite-derived TADR time series acquired in near real time during that eruption. This prediction can be made after an eruption has reached its maximum lava discharge rate and the waning phase of the Wadge trend has begun. This approach therefore only applies to the case of eruption from a chamber undergoing an elastic release of energy during lava flow emplacement, and we provide examples of eruptions where it would not be applicable.

DOI:
10.1007/s00445-017-1134-8

ISSN:
0258-8900