Timescales for permeability reduction and strength recovery in densifying magma

- Porosity-permeability during densification is best described by two power laws.
- The exponent is high and low above and below 0.155 porosity, respectively.
- Large flow channels exist above 0.155 porosity; below, flow is restricted to narrow channels.
- Using these data, we model permeability reduction and strength recovery timescales.
- Results help to understand fracture healing, and eruption and outgassing timescales.

Transitions between effusive and explosive behaviour are routine for many active volcanoes. The permeability of the system, thought to help regulate eruption style, is likely therefore in a state of constant change. Viscous densification of conduit magma during effusive periods, resulting in physical and textural property modifications, may reduce permeability to that preparatory for an explosive eruption. We present here a study designed to estimate timescales of permeability reduction and strength recovery during viscous magma densification by coupling measurements of permeability and strength (using samples from a suite of variably welded, yet compositionally identical, volcanic deposits) with a rheological model for viscous compaction and a micromechanical model, respectively. Bayesian Information Criterion analysis confirms that our porosity-permeability data are best described by two power laws that intersect at a porosity of 0.155 (the “changepoint” porosity). Above and below this changepoint, the permeability-porosity relationship has a power law exponent of 8.8 and 1.0, respectively. Quantitative pore size analysis and micromechanical modelling highlight that the high exponent above the changepoint is due to the closure of wide (∼200-300 μm) inter-granular flow channels during viscous densification and that, below the changepoint, the fluid pathway is restricted to narrow (∼50 μm) channels. The large number of such narrow channels allows porosity loss without considerable permeability reduction, explaining the switch to a lower exponent. Using these data, our modelling predicts a permeability reduction of four orders of magnitude (for volcanically relevant temperatures and depths) and a strength increase of a factor of six on the order of days to weeks. This discrepancy suggests that, while the viscous densification of conduit magma will inhibit outgassing efficiency over time, the regions of the conduit prone to fracturing, such as the margins, will likely persistently re-fracture and keep the conduit margin permeable. The modelling therefore supports the notion that repeated fracture-healing cycles are responsible for the successive low-magnitude earthquakes associated with silicic dome extrusion. Taken together, our results indicate that the transition from effusive to explosive behaviour may rest on the competition between permeability reduction within the conduit and outgassing through fractures at the conduit margin. If the conditions for explosive behaviour are satisfied, the magma densification clock will be reset and the process will start again. The timescales of permeability reduction and strength recovery presented in this study may aid our understanding of the permeability evolution of conduit margin fractures, magma fracture-healing cycles, surface outgassing cycles, and the timescales required for pore pressure augmentation and the initiation of explosive eruptions.