Can we infer the magma overpressure threshold before an eruption? Insights from ground deformation time series and numerical modeling of reservoir failure.

Thursday, 17 December 2015
Poster Hall (Moscone South)
Fabien Albino, University of Miami, Marine Geology and Geophysics, Miami, FL, United States
Overpressure within a magma chamber is a key parameter to understanding the onset of an eruption. Recent investigations indicate that surface inflation at a volcanic edifice does not always precede eruption (Chaussard and Amelung, 2012; Biggs et al., 2014), suggesting that the overpressure threshold may differ between volcanoes. To understand the failure conditions of a magma reservoir, mechanical models were developed to quantify the range of overpressure affordable in a reservoir for a given situation. Even if the choice of the failure criterion is still debated, most investigators agree that the overpressure required to fail the magma reservoir is at first order a function of the crustal stress field and the shape of the magma reservoir. Radar interferometry (InSAR) provides a large dataset of ground deformation worldwide, but many of these InSAR studies continue to use point or dislocation sources (Mogi, Okada) to explain deformation on volcanoes. Even if these simple solutions often fit the data and estimate the depth and the volume change of the source of deformation, key parameters such as the magma overpressure or the mechanical properties of the rocks cannot be derived. We use mechanical numerical models of reservoir failure combined with ground deformation data. It has been observed that volume change before an eruption can easily range one or two order of magnitude from 1-100x106 m3. The first goal of this study is to understand which parameter(s) control the critical volume changes just before the failure of the reservoir. First, a parametric study is performed to quantify the effect of the geometry of the reservoir (radius, depth), the local stress (compressive/extensive) and even the crust rheology (elastic/viscoelastic). We then compare modeling results with several active volcanoes where long time series of volume change are available: Okmok and Westdahl in Alaska, Sinabung and Agung in Indonesia and Galapagos. For each case, the maximum overpressure expected inside the reservoir is estimated and constraints on mechanical parameters of the rocks such as the tensile strength and the shear modulus are provided.