Plain Language Summary
Magma chambers in the continental crust are believed to be “mushy”,
meaning they are reservoirs rich in both crystals and magma. The magma
occupies the pore space between a connected network of crystals and the
difference in density between crystal and magma leads to separation.
During the separation process, the crystal networks behave like a sponge
and magma percolates upward and is extracted as the pore space in the
network of crystals closes. Magma collects atop the “sponge” and can
potentially go on to feed volcanic eruptions at the Earth’s surface.
Therefore, how quickly it can separate has implications for monitoring
volcanic hazards. When the porosity of the “sponge” is sufficiently
low, the closing of pore space can only proceed if individual crystals
are deformed (bent, for instance) and the process is slow. At larger
porosities, however, the pore space can be closed by the sliding or
rotation of crystals. We model this sliding process and compare our
model to analog experiments and find that sliding may allow for this
process to be efficient.
1 Introduction
Forecasting volcanic hazards depends largely on our understanding of how
quickly melt can mobilize and congregate. The physical mechanisms that
lead to phase separation in silicic, crustal magma systems span a
spectrum of melt fractions, ranging from crystal settling at large melt
fractions to compaction (closure of pore space in matrix forming
crystals) at intermediate to low melt fractions (< ca. 0.6 –
0.7) (Bachmann & Bergantz, 2004, Bachmann & Huber, 2019, Holnesset al. , 2017, McKenzie, 1984, Richter & McKenzie, 1984). The
extent to which melt can be effectively separated and the associated
timescale of this process depends on the physical mechanism operating.
Several experimental datasets have been presented by researchers that
aim to constrain the rheology of multiphase magma systems. These
datasets include deformation experiments on partially molten rock
samples, typically exploring melt fractions ranging from ca. 0 – 0.3
(Hirth & Kohlstedt, 1995, Lejeune & Richet, 1995, Mei et al. ,
2002, Renner et al. , 2003, Scott & Kohlstedt, 2006) and can be
used to explore phase separation mechanisms. This suite of experiments
subject partially molten rock samples to applied stresses or controlled
pressure differences and record strain rates as a function of melt
fraction. A combination of stress-strain rate and microstructure
analysis provides insight into the mechanism by which the experimental
samples deform. Over the past two decades, a suite of centrifuge
experiments has also been developed, which can be used to constrain the
rheology of multiphase magma systems (Bagdassarov et al. , 2009,
Connolly et al. , 2009, Krättli & Schmidt, 2021, Manoochehri &
Schmidt, 2014, Schmidt et al. , 2012). One group of these
centrifuge experiments (Connolly et al. , 2009) have been
conducted at melt fractions < 0.3, while another group
(Schmidt et al. , 2012) were conducted at intermediate melt
fractions (ca. 0.3 – 0.7). Connolly and Schmidt (2022) inferred that
the centrifuge experiments were limited by grain boundary-controlled
diffusion (GBD). Microstructural evidence for GBD was reported in this
suite of experiments and included flattened, melt-free grain contacts.
Furthermore, the viscosities inferred by analysis of both sets of
experiments are self-consistent and imply compaction rates much higher
than determined from the suite of partially molten rock experiments
(Fig. 1 ).