Dispersion modeling of the volcanic sulfur dioxide plumes from the simultaneous eruptive activity of Stromboli and Mt Etna on 28 August 2019

During the summer of 2019, both Mt. Etna and Stromboli volcanoes in Sicily were in the stage of no ordinary activity. Mt. Etna was featured by mild strombolian activity from the summit South East Crater producing a moderate SO₂–ash rich plume 4 km above sea level (asl). Meanwhile, at 120 km far from Etna, on 3 July and 28August, the ordinary and typical mild explosive eruptive activity of Stromboli was interrupted by two paroxysms. Both events were characterized by pyroclastic flows and consistent emission of ash–SO₂ rich plume, which spread up to height of 5–6 km asl.
In this work, we explored the spatial dispersion of the volcanic plumes released by both Mt. Etna and Stromboli on August 28 by employing the Weather Research and Forecasting Chemistry (WRF–Chem) model. The simulation was specifically configured and run by considering the time-variable Eruptive Source Parameters (ESPs) related to the SO₂ flux data for Stromboli and Mount Etna observed from ground by the FLAME DOAS scanning spectrometers network.
In order to assess the predictive performance of the WRF–Chem model, the simulated SO₂ dispersion maps were compared with data retrieved on 28 August from TROPOMI and OMI sensors onboard Sentinel–5p and Aura satellites. The results show a good agreement between WRF–Chem and satellite data. In fact, the simulated total mass of the emitted SO₂ from the two volcanoes has the same order of magnitude as the satellite data. However, for the case of Stromboli, the total SO₂ mass predicted by the WRF–Chem simulation is underestimated; this is likely due to inhibition of the real syn-eruptive SO₂ detection by FLAME due to the extreme ash–rich volcanic plume released during the paroxysm.
In conclusion, the results of these two test–cases demonstrate the feasibility of WRF–Chem model with a time-variable ESPs in reproducing different levels of volcanic SO₂ and their dispersion into the atmosphere. For these reasons, our approach could represent an effective support for the assessment of local–to-regional air quality and flight security and, in case of particularly intense events, also on a global scale.