Predicting the End of Lava-Flow-Forming Eruptions from Space

Abstract:

The lava effusion rate, the volumetric flux of lava from the vent, is of great importance in lava flow forecasting and in constraining magma supply conditions. It has been established in the literature that the effusion rates for basaltic lava flows exhibit a typical variation with time, hereafter referred as the “Wadge curve”, with effusion rates rising rapidly to a peak during the very early part of an eruption, before waning over a longer period of time until the eruption ends. We have used thermal infrared remote sensing data acquired by NASA’s MODerate Resolution Imaging Spectroradiometer (MODIS) to estimate effusion rates for 107 eruptions from 35 volcanoes over the last 15 years. Two contrasting methods, those of Harris et al. (1997) and Coppola et al. (2013), were compared. Using the Wadge curves computed from the satellite data for each eruption, we predict (based on the peak effusion rate and a decay constant that describes the exponential waning phase) the time at which effusion rate equals zero, and hence the eruption ends. In order to produce a predictive tool, we derived decay constants for a subset of our eruptions (as training data), and then used these decay constants to predict the end of the remaining eruptions in our dataset. We have determined the decay constant that best describes all the eruptions and then used this global value to retro-cast the end of each eruption in our database to establish how well eruption cessation can be predicted using a globally applicable decay constant. Our approach is of great interest for monitoring of effusive activity and hazard mitigation, via the possibility that by the time an eruption reached its peak effusion rate, as observed from space, we would be able to predict when that eruption would stop.