Page 426 - Caldera Volcanism Analysis, Modelling and Response

Hydrothermal Fluid Circulation and its Effect on Caldera Unrest 401

its reaction to external controls, such as tectonic events. Physical models of
hydrothermal fluid circulation allow the simulation of unrest phenomena related to
the hydrothermal system, and elucidate what controls observed changes in the system.

4.1. Geochemical monitoring
Surveillance programs on active volcanoes commonly involve geochemical
monitoring of hot springs, fumaroles and thermal waters. Hydrothermal fluids
reveal important information about subsurface conditions, and unrest phenomena
are often accompanied by changes in the temperature, composition and discharge
rates of hydrothermal waters (Newhall et al., 2001).
Geochemical monitoring is based on the concept that volcanic gas emissions are
fed by magma degassing and thus reflects to some extent the conditions under
which degassing takes place. The composition of gases exsolving from a magma
chamber will depend on magma composition, the solubility of different gas
components and on the depth and temperature of the magma reservoir. Less soluble
gases, such as nitrogen or carbon dioxide, exsolve first and as degassing proceeds
more soluble species (such as sulphur compounds, water or halogens) will be
progressively released. The ratio of less soluble to soluble components is therefore
expected to change through time, as magma looses its volatiles (Carroll and
Webster, 1994; Delmelle and Stix, 2000). Departure from the expected trends may
indicate a change in the magmatic source, for example by the arrival of a new,
gas-rich magma batch or by the ascent of the degassing magma to shallower depths.
An increase in gas flow rate is also expected if a larger amount of magma is available.
Discharge rate at hot springs increased dramatically before eruptions at Sakurajima
(Japan) in 1914, and at Monte Nuovo (Phlegrean Fields, Italy) in 1538 (Newhall
and Dzurisin, 1988). Mount Usu (Japan) eruption in 2000 was preceded by an
increase in carbon dioxide degassing within the summit caldera (Herna ´ndez et al.,
2001). Major diffuse degassing of carbon dioxide, killing trees in Mammoth
Mountain area, was also observed during unrest at Long Valley caldera, California
(Pribnow et al., 2003; Bergfeld et al., 2006). Seismicity and ground deformation
have been recorded there since 1980, and this has been accompanied by a recorded
increase of well fluid pressure and by a higher flow rate of magmatic gases (Sorey et al.,
2003, and reference therein). The interpretation of monitoring data is, however,
not always straightforward. Volcanic gas emissions do not simply reflect the process
of magma degassing. As magmatic gases rise toward the surface they are affected by
cooling, decompression, oxidation and reactions with host rock and groundwaters.
Fluids sampled at the surface result from complex interactions between the deep,
magmatic contributions and shallower components of the hydrothermal system. As
hot fluids interact with shallow water bodies, heat exchange, phase transition and
chemical reactions may take place: ascending fluids may gain or loose water vapour,
depending on whether condensation or evaporation prevails. In addition, reactive
and soluble components, such as sulphur dioxide or halogens, may be lost in
groundwaters and appear substantially depleted in volcanic gas emissions. Sulphur
compounds are particularly sensitive to secondary processes and redox conditions,
which ultimately control the SO /H 2 S ratio; deposition (or revolatilisation) of
2