TOWARDS A GLOBALLY CONSISTENT DYNAMIC PICTURE OF PRE-ERUPTION VOLCANO DYNAMICS

Understanding the processes and dynamics characterizing underground magmatic bodies, and their relationships with recorded geophysical signals, is necessary for the implementation of robust early warning systems and short-term volcanic hazard forecast. That understanding requires a multi-disciplinary approach where virtually all branches of volcanology concur to the definition of a consistent and physically sound dynamic picture of the volcanic system. Past eruptive histories, magma and fluid geochemistry, petrology, and geophysical/geochemical surveys provide constraints to the underground system in terms of size, geometries, presence of multiple reservoirs, composition of magmas involved, volatile species, characteristics of country rocks, etc. Laboratory determinations and modeling of relevant magma/rock properties provide constitutive equations at the P-T conditions encountered up to several km in the crust, together with knowledge of the basic physics characterizing Earth materials at different dynamic states. Physical and mathematical modeling of transport phenomena in multiphase fluids with phase changes, and numerical solution of the resulting system of non-linear partial differential equations, allow a time-space description of the processes characterizing movements of magmas in deep volcanic regions, including mixing of compositionally diverse magmas. Integration of time-space mass distribution, and dynamic coupling with rock elasto-dynamics, leads finally to the determination of micro-gravity anomalies and ground displacement dynamics in the quasi-static and seismic frequency bands. First applications to real volcanoes covering a range of magma compositions, volatile contents, size and geometrical complexity of the magmatic system, allow an evaluation of the different roles of several quantities in determining magma dynamics and associated signals. All of the results obtained so far concur to suggest that i) simple point-source or homogeneous source assumptions commonly employed in signal inversion analyses can lead to substantial misinterpretations, and that ii) correlated ground displacement and gravity oscillations in the Ultra-Long-Period frequency range (> or >> 100 s) can be diagnostic of on-going magma convection dynamics.