Cyclicity at active volcanoes is commonly observed and interpreted to reflect periods of vesiculation and gas pressurisation beneath more dense degassed plugs. In 1997, Vulcanian eruptions from Soufrière Hills volcano (SHV), Montserrat, had a mean repose time of 9.5 hours. We consider the scenario that following each eruption a ~9-hour cycle of pore-collapse and resultant gas-phase pressurisation occurred during continued magma degassing and variable outgassing. This process stores energy sufficient for fragmentation. To constrain such a scenario, samples of andesite pumice with initial porosities of 60-65% and 75-77% were heated at 2°C.min-1 to isotherms of 900°C and 940°C under atmospheric pressure for 0.5-10 hours. We measured the change in the pore network by helium pycnometry, P and S wave velocities and permeability. We used X-ray computed tomography and optical microscopy to compare the textural evolution of experimentally deformed samples with naturally deformed and dense samples from the same eruption. Our results suggest that magmatic foam collapse under these conditions is primarily controlled by the viscosity and is therefore temperature and time dependent. Porosity and permeability decrease on timescales and at temperatures relevant to SHV cyclic eruptions. The time required for permeability to decrease significantly depends on the starting pore network geometry of the magma following the previous explosive eruption. Samples at 900°C and 940°C for 2 hours lost averages of 6% and 12% connected porosity, respectively, compared with 30% and 45% after 9 hours. We propose that the onset of dense plug formation can occur by magma foam collapse and that this is a primary mechanism for pore-pressurisation leading to magma fragmentation during cyclic Vulcanian eruptions. As the densified top of the magma column thickens, load from magma overburden will accelerate the process we describe. Because permeability is the factor controlling the response time of bubbly magma to pressure changes, our study has implications for predicting the repose time of cyclic eruptions.
Cite this work
Researchers should cite this work as follows: