Jökull - 01.12.1985, Qupperneq 3

The Petrologic Estimation of Volcanic Degassing
H. SIGURDSSON, J.D. DEVINE and A.N. DAVIS
Graduate School of Oceanography
University of Rhode Island
Kingston, RI 02881 U.S.A.
ABSTRACT
Petrologic estimates of degassing of sulfur, chlorine
and fluorine and of the total volcanic volatile mass from
several eruptions are closely comparable to volcanic
aerosol estimates based on total acidity in ice-cores and
direct stratospheric satellite-based extinction measure-
ments. The results of petrologic studies of twenty erup-
tions indicate that three general types of volcanic degas-
sing should be considered: (a) dominated by sulfuric acid
from degssing of magma, (b) dominated by halogens
(chlorine and lesser fluorine) from degassing of magma,
and (c) formed by decomposition of sulfur-rich mineral
phase e.g. anhydrite during eruption (El Chichon-type).
INTRODUCTION
Explosive volcanic eruptions inject significant quanti-
ties of tephra, consisting of silicate glass, crystals and
lithic fragments, and volcanic gases into the earth’s
atmosphere. The residence time of most of the tephra is
relatively short, however, due to its large diameter and
the process of particle aggregation (Carey and
Sigurdsson 1982). Some volcanic gases (e.g. S02, H2S)
on the other hand, undergo gas to particle conversion
by oxidation and reaction with atmospheric water
(Arnold and Buhrke 1983) and thus may form a stratos-
pheric volcanic aerosol layer of global extent following a
large eruption. Major volcanic aerosols have also been
produced following massive lava eruptions, even though
the production of tephra was minor (Sigurdsson 1982a).
Because of the potential climatic impact of volcanic
aerosols (Hansen et al. 1981) it is clear that their study is
of great importance, in particular the determination of
the mass and composition of volcanic components of the
stratospheric aerosol layer. The first satellite-based
study of volcanic aerosol mass loading was undertaken
in 1979 during the eight small explosive eruptions of
Soufriere volcano, St. Vincent. The total mass of stra-
tospheric ejecta from plumes of only two of these
eruptions (13, 14 April and 17 April) was found to be
2.3X109 g (McCormick et al. 1981). Similar methods
were applied to the study of the Mount St. Helens 1980
volcanic aerosol, where the total mass was determined
as 3xl0n g (Kent 1982, McCormick 1982). A variety of
techniques have been applied recently in determining
volcanic mass loading of the stratospheric aerosol layer
emitted from the 1982 eruption of E1 Chichon volcano
in Mexico. They include airborne lidar observations of
backscattering ratio (McCormick and Swissler 1983)
and balloon-borne particle counters (Hofmann and
Rosen 1983) which indicate a stratospheric aerosol mass
loading of 1—2xl013 g. We have previously shown that
the large sulfur release from this eruption can be
accounted for by decomposition of anhydrite (CaS04)
phenocrysts found in the tephra (Devine et al. 1984).
Prior to 1979 direct measurements of the mass of
volcanic aerosols injected into the stratosphere from
specific eruptions were generally not feasible. Instead,
methods were developed to estimate the mass of vol-
canic aerosols from changes in the acidity of Greenland
ice cores (Hammer 1977, 1980) and changes in stratos-
pheric aerosol optical depth (Stothers 1984). We have
recently used a method for the determination of vol-
canic volatile release (Sigurdsson 1982a, Devine et al.
1984) based on petrologic analysis of samples of the
quenched magmas from the eruptions of interest. The
method which is based on estimation of pre-eruption
volatile content obtained by analysis of glass inclusions
in phenocrysts, has the advantage that the volcanic
volatiles from ancient but climatologically significant
eruptions can be studied, and that estimates of mass and
composition of the volcanic volatile can be made from
sampies of a well-dated deposit from specific volcano,
whereas the other mothods of observation must specu-
late about the source and, in the case of ice-cores, the
timing of volcanism. Rampino and Self( 1984) claim that
our petrologic and volcanological estimates of volatile
yield are underestimates compared with data based on
optical depth and ice-core acidity. As shown below,
their assertion is unfounded, and is based on erroneous
interpretation of ice-core data. It is in fact evident from
our published results (Devine et al. 1984) that the ice-
core and petrologic estimates of volcanic volatile emis-
sion from the e.g., two largest studied events (the Laki
1783 and Tambora 1815 eruptions) are indeed remark-
JÖKULL 35. ÁR 1