Near Real Time Forecasting of Lava Flow Paths Using Magflow Modeldriven by Thermal Satellite Data

Timely predictions of the areas likely to be inundated by lava flows are of major interest to hazard managers during a volcanic eruption. To this aim several models have been developed, which deal with various aspects of flow emplacement and cooling. That challenge has inspired the INGV-CT to develop the MAGFLOW Cellular Automata model to simulate lava flows [6]. It is based on a steady state solution of the Navier-Stokes equations coupled with a heat transfer model that takes into account radiative heat losses [1, 4]. The effects of core cooling on flow rheology are modeled via the temperaturedependent viscosity model, derived for Etna. MAGFLOW was applied with satisfactory results to simulate flow fields formed during the 2001 and 2004 eruptions at Mt Etna, Sicily [1, 4, 6]. For a given composition, the volumetric flux of lava from the vent (i.e. the lava discharge rate) is the principal parameter controlling final flow dimensions [3]. Direct field measurements of lava discharge rate can be made. However, in most situations, measurements, especially regular measurements, are difficult-to-impossible due to safety and logistical reasons. For example, during the early phases of an eruption, lava flows spread at high velocities, and explosions, fountaining and lava spattering may occur at the master vent, making close approach to the master channel hazardous and discharge rate measurement unsafe. The timely and synoptic view afforded by satellite-based sensors can be used to estimate time-averaged discharge rate throughout an eruption [2,4]. To this end, we have developed a software tool that uses near-real-time infrared satellite data acquired by different sensors (MODIS and SEVIRI) to estimate discharge rates. These time-varying discharge rates are then used to drive MAGFLOW simulations to chart the spread of lava as a function of time. We tested our thermal monitoring system (Fig. 1) on Etna volcano during the first 40 days of May 2008 eruption. The good agreement between simulated and mapped flow areas indicates that model-based inundation predictions, driven by time-varying discharge rate data, provide an excellent means for assessing the hazard posed by on-going effusive eruptions.