Page 48 - Caldera Volcanism Analysis, Modelling and Response

Residence Times of Silicic Magmas Associated with Calderas 23

Figure 7 Very simpli¢ed map showing LongValley caldera (topographic margin) and the
main units discussed in the text. BT, BishopTu¡, GM, Glass Mountain (pre-caldera). Figure
based on Bailey (1989) and on a simpli¢ed map published on the USGS web page.

0.3 My for the younger lavas, and up to 0.7 My for the old lavas. Halliday et al.
(1989) proposed that high Rb/Sr values and low Sr concentrations of the lavas were
due to extreme crystal fractionation and the isochrons were dating such a process.
The geochemical similarity of the younger Glass Mountain lavas and the Bishop
Tuff led them to suggest that the chamber containing the magma later to erupt as
the Bishop Tuff was already formed, including the chemical zoning, by about
1.1 Ma. This implied the existence of relatively small volumes of mostly liquid
magma at shallow depths for long periods of time. Sparks et al. (1990) raised
questions on the Halliday et al. (1989) interpretation mainly related to the feasibility
of keeping such volumes of magma in a mostly liquid state without freezing in the
upper crust. Their alternative interpretation was that the silicic magmas were the
result of remelting of a granite source region (presumably in deep portions of
the crust) where they develop the ‘old isochrons’ but the residence times in
the shallow crust could be short. In a new model by Mahood (1990) the volumes
of erupted rhyolites were small compared to the size of the entire reservoir,
minimising the thermal problems of keeping small and mainly liquid magma
batches for a long time in the shallow crust. Moreover, Mahood (1990) argued that
parts of the reservoir can be ‘frozen’ as crystallised rind or immobile mush, and a
portion can be ‘defrosted’ for eruption at a later date by input of new magma
without causing the magma to be shifted off the isochron. Thus, the long residence
times could be explained if the erupted magmas were solidified in the upper crust
for some time and later remelted close to eruption.
Davies et al. (1994) and Davies and Halliday (1998) reported more precise
40 39
Ar/ Ar and Rb–Sr and Sm–Nd isotope data, and the residence times for the
Glass Mountain magmas were reduced to about half that originally proposed,
with a maximum time of ca. 350 ky (Table 6). Moreover, Davies et al. (1994) and