Long-term geochemical surveillance of fumaroles at Showa-Shinzan dome, Usu volcano, Japan

Abstract This study investigates 31 years of fumarole gas and condensate (trace elements) data from Showa-Shinzan, a dacitic dome-cryptodome complex that formed during the 1943–1945 eruption of Usu volcano. Forty-two gas samples were collected from the highest-temperature fumarole, named A-1, from 1954 (800 °C) to 1985 (336 °C), and from lower-temperature vents. Condensates were collected contemporaneously with the gas samples, and we reanalyzed ten of these samples, mostly from the A-1 vent, for 32 cations and three anions. Modeling using the thermochemical equilibrium program, SOLVGAS, shows that the gas samples are mild disequilibrium mixtures because they: (a) contain unequilibrated sedimentary CH 4 and NH 3 ; (b) have unequilibrated meteoric water; or (c) lost CO, either by air oxidation or by absorption by the sodium hydroxide sampling solution. SOLVGAS also enabled us to restore the samples by removing these disequilibrium effects, and to estimate their equilibrium oxygen fugacities and amounts of S 2 and CH 4 . The restored compositions contain > 98% H 2 O with minor to trace amounts of CO 2 , H 2 , HCl, SO 2 , HF, H 2 S, CO, S 2 and CH 4 . We used the restored gas and condensate data to test the hypotheses that these time-series compositional data from the dome's fumaroles provide: (1) sufficient major-gas data to analyze long-term degassing trends of the dome's magma-hydrothermal system without the influence of sampling or contamination effects; (2) independent oxygen fugacity-versus-temperature estimates of the Showa-Shinzan dacite; (3) the order of release of trace elements, especially metals, from magma; and (4) useful information for assessing volcanic hazards. The 1954–1985 restored A-1 gas compositions confirm the first hypothesis because they are sufficient to reveal three long-term degassing trends: (1) they became increasingly H 2 O-rich with time due to the progressive influx of meteoric water into the dome; (2) their C S and S Cl ratios decreased dramatically while their Cl F ratios stayed roughly constant, indicating the progressive outgassing of less soluble components (F ≈ Cl > S > C) from the magma reservoir; and (3) their H 2 O H 2 , CO 2 CO and H 2 S SO 2 ratios increased significantly in concert with equilibrium changes expected for the ~ 500 °C temperature drop. When plotted against reciprocal temperature, the restored-gas log oxygen fugacities follow a tight linear trend from 800 °C to NNO +2.5 at ~ 400 °C. This trend largely disproves the second hypothesis because the oxygen fugacities for the The time-series trace-element data also disprove the third hypothesis because rock- and incrustation-particle contaminants in the condensates account for most of the trace-element variation. Nonetheless, highly volatile elements like B and As are relatively unaffected by this particle contamination, and they show similar time-series trends as Cl and F. Finally, except for infrequent sampling around the 1977 Usu eruption, the results generally confirm the fourth hypothesis, since the time-series trends for the major gases and selected trace elements indicate that, with time, the system cooled, degassed and was infiltrated by meteoric water, all of which are positive signs that volcanic activity declined over the 31-year history. This study also suggests that second boiling of shallow magma within and possibly beneath the cryptodome sustained magmatic degassing for at least 20 years after emplacement.