A Physical-Experimental Model for Small-Scale Basaltic Vulcanian Eruptions

In the last period of its summer 2001 flank activity Mt. Etna produced ash explosions not common at this basaltic volcano. The explosions took place at a new vent at 2550 m.a.s.l. and followed Strombolian and effusive activity. At first the ash erupted as a continuous, pulsing plume a few km high, occasionally undergoing small-scale, partial collapses. Afterward the frequency and intensity of the explosions decreased to isolated events, gradually fading to an end. Each explosion produced mostly wind-driven ash and subordinate blocks ballistically emplaced close to the vent. The final deposit is an ash layer 20 cm thick at 50 m from the vent. Both the ash and the blocks are poorly vesicular with a microcrystalline matrix. This type of activity, deposit, and products closely match, although on a smaller scale, those of Vulcanian explosions: hence we hypothesize that a similar mechanism of overpressurization of a magma plug was at work at Etna. Petrological observations indicate that the plug formed by gradual crystallization of the stagnant magma at the top and at the borders of the conduit after the Strombolian and effusive activity. In order to model the ash explosions we applied three independent techniques to estimate the overpressure in the exploding plug: 1) HP-HT shock tube experiments to determine the pressure differential required to fragment the plug material; 2) calculation of the overpressure generated by plug crystallization; 3) application of two physical models for vulcanian eruptions using the range of ballistic blocks.

To measure the minimum pressure differential required to fragment them, we heat and pressurize cylindrical cores from the blocks at T up to 800-900° C and P up to 25 MPa inside a shock-tube apparatus, and then suddenly decompress them to ambient conditions. The pressure threshold varies strongly with sample porosity, from 22.5 MPa at 4% porosity to less that 5 MPa at 20%. Since these values came from the dense blocks that were not fragmented to ash, we assume the lower value as representative of the ash-forming magma.

The ash from the explosions has 35 vol.% more microlites than that from Strombolian activity. Applying available models of crystallization-induced overpressurization and find that crystallization of 35 vol.% of microlites is enough to pressurize a basaltic magma to 4-6 MPa at a depth of 100 m ca..

Finally we use the range of ballistic blocks to estimate their exit velocity that, in turn, can be related to the pressure in the conduit. Using realistic values for the ejection angle, shape and density of the blocks, and other ambient conditions, the maximum range of 450 m for blocks of 60 cm corresponds to a muzzle velocity of 80 ms^-1. Using this value as an input, two different models for vulcanian eruptions give an initial pressure of 4-6 MPa in the assumption that the volatile content of magma in the plug is in the order of 0.1 wt.%. This assumption is in agreement with petrological evidence that the plug formed by the accumulation of largely degassed magma.

In our model, crystallization of 35 vol.% microlites in a magma with 0.1 wt.% of volatiles at a depth of 100 m increase the overpressure in the plug up to 4-6 MPa. At this point the fragmentation threshold is reached and an explosion occurs, ejecting blocks at speeds up to 80 ms^-1. Notwithstanding the fact that other processes may contribute pressurizing the plug, we believe that the convergence of three independent methods under reasonable assumptions strongly supports our model.