Observation of SO 2 degassing at Stromboli volcano using a hyperspectral thermal infrared imager
Résumé
Thermal infrared (TIR) imaging is a common tool for the monitoring of volcanic activity. Broadband cameras with increasing sampling frequency give great insight into the physical processes taking place during effusive and explosive event, while Fourier transform infrared (FTIR) methods provide high resolution spectral information used to assess the composition of volcanic gases but are often limited to a single point of interest. Continuing developments in detector technology have given rise to a new class of hyperspectral imagers combining the advantages of both approaches. In this work, we present the results of our observations of volcanic activity at Stromboli volcano with a ground-based imager, the Telops Hyper-Cam LW, when used to detect emissions of sulfur dioxide (SO2) produced at the vent, with data acquired at Stromboli volcano (Italy) in early October of 2015. We have developed an innovative technique based on a curve-fitting algorithm to quickly extract spectral information from high-resolution datasets, allowing fast and reliable identification of SO2. We show in particular that weak SO2 emissions, such as inter-eruptive gas puffing, can be easily detected using this technology, even with poor weather conditions during acquisition (e.g., high relative humidity, presence of fog and/or ash). Then, artificially reducing the spectral resolution of the instrument, we recreated a variety of commonly used multispectral configurations to examine the efficiency of four qualitative SO2 indicators based on simple Brightness Temperature Difference (BTD). Our results show that quickly changing conditions at the vent – including but not limited to the presence of summit fog – render the establishment of meaningful thresholds for BTD indicators difficult. Building on those results, we propose recommendations on the use of multispectral imaging for SO2 monitoring and routine measurements from ground-based instruments.