S. Vergniolle, BUBBLE-SIZE DISTRIBUTION IN MAGMA CHAMBERS AND DYNAMICS OF BASALTIC ERUPTIONS, Earth and planetary science letters, 140(1-4), 1996, pp. 269-279
In the shallow magma chambers of volcanoes, the CO2 content of most ba
saltic melts is above the solubility limit. This implies that the cham
ber contains gas bubbles, which rise through the magma and expand. Thu
s, the volume of the chamber, its gas volume fraction and the gas flux
into the conduit change with time in a systematic manner as a functio
n of the size and number of gas bubbles. Changes in gas flux and gas v
olume are calculated for a bubble size distribution and related to cha
nges in eruption regimes. Fire fountain activity, only present during
the first quarter of the eruption, requires that the bubbles are large
r than a certain size, which depends on the gas flux and on the bubble
content [1]. As the chamber degasses, it loses its largest gas bubble
s and the gas flux decreases, eventually suppressing the fire fountain
ing activity. Ultimately, an eruption stops when the chamber contains
only a few tiny bubbles. More generally, the evolution of basaltic eru
ptions is governed by a dimensionless number, tau = tau g Delta rho a
(o)(2)/(18 mu h(c)), where tau = a characteristic time for degassing;
a(o) = the initial bubble diameter; mu = the magma viscosity; and h(c)
= the thickness of the degassing layer. Two eruptions of the Kilauea
volcano, Mauna Ulu (1969-1971) and Puu O'o (1983-present), provide dat
a on erupted gas volume and the inflation rate of the edifice, which h
elp constrain the spatial distribution of bubbles in the magma chamber
: bubbles come mainly from the bottom of the reservoir, either by in s
itu nucleation long before the eruption or within a vesiculated liquid
. Although the gas flux at the roof of the chamber takes similar value
s for both eruptions, the duration of both the fire fountaining activi
ty and the entire eruption was 6 times shorter at Mauna Ulu than durin
g the Puu O'o eruption. The dimensionless analysis explains the differ
ence by a degassing layer 6 times thinner in the former than the latte
r, due to a 2 year delay in starting the Mauna Ulu eruption compared t
o the Puu O'o eruption.