Magma intrusion and deformation predictions: Sensitivities to the Mogi assumptions

作者：Masterlark Timothy^*

来源：JOURNAL OF GEOPHYSICAL RESEARCH, 2007, 112(B6): B06419.

DOI：10.1029/2006JB004860

Mogi's (1958) magma intrusion model is widely used to predict observed deformation of active volcanoes. The model simulates a small spherical expansion source (SES) embedded in a homogeneous, isotropic, Poisson-solid half-space (HIPSHS). This study computes surface displacement due to SESs at depth using a combination of analytical and finite element models (FEMs), for which the HIPSHS assumptions are not required. Interferometric synthetic aperture radar (InSAR) data suggest that Okmok volcano, Alaska, subsided more than a meter owing to lava extrusion during its 1997 eruption. Inverse methods, which use an HIPSHS model, precisely locate an SES at a depth of 3100 m beneath the center of the caldera. A series of alternative model configurations relax the combined suite of HIPSHS assumptions and sequentially isolate the effects of each assumption. Forward modeling predictions are relatively insensitive to topographic effects and layered elastic properties, somewhat sensitive to anisotropic elastic properties and Poisson's ratio, and very sensitive to the presence of weak materials within a caldera. Inverse methods, combined with analytical and FEM-generated impulse response functions, isolate the influence of each HIPSHS assumption on SES depth and pressure estimations. Results are particularly sensitive to a model configuration simulating a weak caldera, for which the estimated SES depth (4500 m) is significantly deeper than the estimate for the HIPSHS model. For deformation data having high signal-to-noise ratios, such as the co-eruption InSAR data for Okmok volcano, both forward and inverse deformation prediction errors attributed to the Mogi (1958) assumptions can greatly exceed observation uncertainties.

• 出版日期2007-6-26