Listening to Ambrym Volcano (Vanuatu) By an Acoustic Array: Cyclictity in Gas Volume at an Open Vent Volcano

Tuesday, 16 December 2014
Sylvie Vergniolle1, Christelle Zielinski1, Philipson Bani2, Michel Lardy3, Alexis LE Pichon1,4, Philippe Millier4, Pascal Herry4, Sylvain Todman5 and Esline Garaebiti5, (1)Institut de Physique du Globe de Paris, Paris, France, (2)Laboratoire Magma et volcans, Clermont-Ferrand, France, (3)Institut de Recherche pour le Developpement, Noumea, New Caledonia, (4)Commissariat a l’Energie Atomique, Bruyeres le chatel, France, (5)Vanuatu Meteorology and Geohazards Department, Port Vila, Vanuatu
Volcanic activity at Ambrym volcano (Vanuatu) is currently characterised by continuous lava lake activity within two main cones (Marum and Benbow), distant by 3 km. Acoustic measurements were performed by an acoustic array set at a distance of 3 km from the active vents between April 2009 and January 2010. The detection of volcanic signals, done by using PMCC (progressive multi-channel correlation), showed that more than 97 % of the detections came from one of the active vents, with millions of acoustic transients and no quiescent period. Two short Strombolian phases and 3 isolated explosions existed at Marum during this period while no strong signal was detected at Benbow. The active degassing, i.e. that detected by acoustic measurements, is mostly around 1 Hz at both vents. The gas volume is estimated from calculating acoustic power for each detection in the 0.8-2 Hz frequency range, assuming that the source has a constant radius of 5 m. The gas volume expelled per hour is similar at Benbow and Marum with 0.5-4 Mm^3. One striking observation is the existence of regular alternations between periods with high and small expelled gas volume, with 3 Mm^3 per hour and 0.5 Mm^3 per hour for each period, respectively. The detection of acoustic pulses even small, which is done with wavelet analysis, showed rapid and regular variations in the number of acoustic transients (1-460). The duration, return time and expelled gas volume were similar at Benbow and Marum on the 50 and 38 recognised cycles, respectively. The duration, return time and expelled gas volume at Benbow were equal to 58 hr, 70 hr and 42 Mm^3, respectively. Their relative small standart deviation for the above parameters, equal to 24 hr, 82 hr and 26 Mm^3 respectively, suggested that the fluctuations between periods with high and low gas volume, is probably cyclic. This 5 days-cyclicity could be perhaps be related to thermal convection processes in the magma reservoir.