Page 61 - Caldera Volcanism Analysis, Modelling and Response

36 Fidel Costa

Figure 9 Calculated residence times vs. volume for all deposits.The two curves are
solidi¢cation models calculated using Equation (3) and the following parameters: (1) an initial
temperature of 9001C for the magma.This is 725 degrees of the liquidi of most magmas treated
2 1
in this manuscript, (2) a thermal di¡usivity of 10 6 m s (e.g., Gue ¤ guen and Palciauskas,
1994), (3) a solidus of 6501C, which is representative for silicic and water bearing systems (e.g.,
Johannes and Holtz,1996), and (4) two di¡erent conditions for the initial temperature of the host
rock. In one case a temperature of 2001C was used as representative of a depth of 6 km (curve
that reaches solidi¢cation faster). In the other, the temperature of the host rock increased with
1
increasing size of the reservoir at typical geothermal gradient of 30 K km , starting from
3
2001Cat 6 km fora 1km body. As can be seen from the ¢gure the e¡ect of incorporating a
3
geothermal gradient does not signi¢cantly change the results. Most magmasoca.100 km lie
to the left of the solidi¢cation curves suggesting their residence times re£ect probably crystal
recycling from previous igneous activity (although see Section 4.1.1) rather than the time they
spend above the solidus.
Oruanui and Rotoiti magmas at 4 and 9 ky, respectively (both from Taupo volcanic
zone), followed by the Dry Creek and Lava Creek Yellowstone magmas at 7
and 19 ky, respectively. There is not a good correlation between the volume and
3
residence time, but most eruptions o10 km have residence times o100 ky, and
3
those W100 km have longer residence times, up to ca. 400 ky for the Fish Canyon
and W500 ky for La Pacana. A positive correlation between the volume of the
deposit and the residence time can be expected if magma fluxes of small and large
systems are the same (see magma production rate section below). The question is
what these magma residence times indicate. If we first exclude the possibilities of
(1) the effect of different closure temperatures consequent upon the use of different
phases and radioactive systems, (2) the obvious xenocrysts which are several My
older than the host and (3) the possible effects of sample bias due to the analytical
method used (in situ vs. bulk mineral separates), at least two additional end-member
possibilities remain. The residence times could constrain the duration that the