SO2 flux monitoring at Stromboli with the new permanent INGV SO2 camera system: A comparison with the FLAME network and seismological data

We installed a permanent SO2 camera system on Stromboli, Italy, in May 2013, in order to improve our capacity to monitor the SO2 emissions from this volcano. The camera collects images of SO2 concentrations with a period of ~ 10 s, allowing quantification of short-term processes, such as the gas released during the frequent explosions which are synonymous with Stromboli. It also allows quantification of the quiescent gas flux, and therefore comparison with the FLAME network of scanning ultraviolet spectrometers previously installed on the island. Analysis of results from the SO2 camera demonstrated a good agreement with the FLAME network when the plume was blown fully into the field of view of the camera. Permanent volcano monitoring with SO2 cameras is still very much in its infancy, and therefore this finding is a significant step in the use of such cameras for monitoring, whilst also highlighting the requirement of a favourable wind direction and strength.

We found that the explosion gas emissions are correlated with seismic events which have a very long period component. There is a variable time lag between event onset time and the increase in gas flux observed by the camera as the explosion gas advects into the field of view of the camera. This variable lag is related to the plume direction, as shown by comparison with the plume location detected with the FLAME network. The correlation between explosion gas emissions and seismic signal amplitude show is consistent with a gas slug-driven mechanism for seismic event production. Comparison of the SO2 camera measurements of the quiescent gas flux shows a fair quantitative agreement with the SO2 flux measured with the FLAME network. Overall, the SO2 camera complements the FLAME network well, as it allows frequent quantification of the explosion gas flux produced by Stromboli, whose signal is in general too brief to be measured with the FLAME network. Further work is required, however, to fully automate the calculation of SO2 flux from the SO2 images captured with the camera, and to adequately account for scattering effects.