Intercomparison of geochemical techniques at La Soufrière de Guadeloupe (FWI) volcano: their advantages and their limits over a long-standing unrest

Roberto Moretti (1,2), Séverine Moune (1,2), Vincent Robert (1,2), David E. Jessop (1,2), Tristan Didier (1,2), Magali Bonifacie (1,2) & Jean-Christophe Komorowski (1)
(1) Université de Paris, Institut de Physique du Globe de Paris, UMR 7154 CNRS, Paris (F). (2) Observatoire Volcanologique et Sismologique de Guadeloupe, Institut de Physique du Globe de Paris, Gourbeyre (F). (3) Université Clermont Auvergne, CNRS, IRD, OPGC, Labora- toire Magmas et Volcans, Clermont-Ferrand (F). Corresponding author e-mail: moretti@ipgp.fr

Volcanic gas emissions from the main low-temperature fumaroles (around 100°C) of one of the most dangerous volcanoes in the Lesser Antilles, La Soufrière de Guadeloupe, are measured routinely using direct sampling with NaOH bottles (for H₂O, CO2, H₂S_tot, H₂, CH₄, CO, N₂, He, Ar, O₂), with P₂O₅ bottles (for H₂O, CO₂, H₂S, SO₂, H) and MULTIGAS (for HO, CO HS, SO, H). This allows us to perform a study about an intercomparison of volcanic gas monitoring methologies. The results show some good overlap and similar temporal evolution between the dataset of the different methologies, which is of utmost of importance for monitoring purposes. However, there is no good agreement between the three methodologies for the same chemical ratio. The differences could be explained by a time- dependant diffusive phenomenon that affects the concentrations of sampled gases for two methods of the three (P₂O₅ bottles and MULTIGAS). P₂O₅-based analyses are affected because effusion on rapid decompression and fast sampling affects the gas volume in the pipeline connecting the vent outlet to the sampling. Although MULTIGAS is a continuous, flow-through, sampling method, results are likely affected because of the sampling architecture involving filter and pump. Moreover, sulphur reaction (precipitation of elemental S upon exit from the fumarole, conversion of H₂S to SO₂, chemical reaction with the P₂O₅ powder that release H₂S) and water condensation can also explain some discrepancies.

unrest, fumarole sampling, effusion/diffusion, NaOH, P₂O₅, MULTIGAS.