Volcanic-gas monitoring - Volcanism and Global Environmental Change - page 89

1994 ) , and for verifying magma convection and passive gas-melt separation in
conduits as likely causes of excess sulfur degassing (Shinohara, 2008 ). As obser-
vations increased in number, the first calculations of the global volcanic sulfur

output were proposed, and there is now general consensus that present-day annual
SO 2 emissions from volcanoes amount to 13 - 18 Tg annually (Mather and Pyle,
this volume).
There is now considerable evidence of precursor variations in SO 2 fluxes being

detectable prior to the eruption of intermediate to silicic volcanoes. Silicic volca-
noes, which frequently undergo periods of dormancy lasting centuries or even
millennia, generally comprise low-temperature (H 2 S-dominated) fumarolic
systems, which constitute the surface expression of subsurface hydrothermal
systems (Giggenbach, 1996 ). Pre-eruptive degassing of fresh magma commonly
leads to hydrothermal fluids being replaced by hotter, more oxidised magmatic

fluids that are rich in SO 2 . This was evident in the evolution of the gas chemical
composition at Soufrière Hills volcano on Montserrat prior to and during the
emplacement of the 1995 dome (Hammouya et al ., 1998 ) and at the re-awaking
of Turrialba in Costa Rica (Vaselli et al ., 2010 ). This process has also been detected
in the SO 2 flux record, such as at Pinatubo in 1991, where the cataclysmal plinian

eruption was preceded by the SO 2 flux increasing by an order of magnitude over a

period of 2 weeks, indicating the intrusion of pre-eruptive magma at shallow depth

(Daag et al ., 1996 ). Improved networks for measuring SO 2

~

fluxes are now being
used to increase our understanding of the transition from quiescence to eruption.
There have been recent reports of long-term and large changes in SO 2 degassing
prior to small phreatic eruptions, such as at Santa Ana (El Salvador) and Turrialba
(Costa Rica) (Olmos et al ., 2007 ;Conde et al ., 2013 ). During long-lived dome
eruptions, degassing trends have shown particularly large SO 2

fluctuations, re ecting

variable rates of ma c magma recharge and degassing (Christopher et al ., 2010 ),

and/or changes in conduit/dome permeability (Stix et al ., 1993 ;Fischer et al ., 1994 ;
Edmonds et al ., 2003 ). At ma c volcanoes, such as Etna (Caltabiano et al ., 1994 )

and Kilauea (Sutton et al ., 2001 ), SO 2 fluxes are strongly correlated with the rates

of shallow magma supply and degassing (Allard et al ., 1994 ) , and ultimately with
the magma extrusion rate (Burton et al ., 2009 ). Exceptionally high SO 2 emissions
have been observed during paroxysmal eruptions, while precursor SO 2 - ux vari-

ations prior to eruptions are more dif cult to identify, possibly due to masking

by large and persistent SO 2 emissions during quiescent phases.

6.4.2 CO 2
The rates of CO 2 release from subaerial volcanism are even less well known
than those of SO 2 .Burton et al .( 2013 ) compiled a list of only 33 volcanoes