What is general behavior and what are system or location specific characteristics during strombolian explosions
Final Report Abstract
Strombolian eruptions at various volcanoes exhibit a wide range of eruption characteristics, and the question to be addressed in this project was to identify common features of Strombolian eruptions as well the development of a model for conduit transport processes during Strombolian eruptions. In order to identify common features during Strombolian eruptions we used data recorded at three different volcanoes (Yasur volcano, Vanuatu; Mt. Erebus, Antarctica; Stromboli volcano, Italy) as well as data recorded during artificial volcanic eruptions to constrain common behavior. Data from Yasur and Mt. Erebus were available and data from Stromboli volcano were recorded in the framework of this project. While recording the data at Stromboli, we used a modified Doppler radar system (this was applied for in the proposal) that allowed for simultaneous measurement of high resolution distance measurements as well as eruption velocities in order to track deformation processes at the vent prior to an eruption. In 2010 we installed the infrastructure (two solar stations) to continuously record data at Stromboli in two locations. In addition the modified Doppler Radar system was installed. Due to technical problems only the data recorded during a second campaign in 2011 could be used for evaluation. A comparison of the four data sets shows that the artificial volcanic eruptions as well as the the eruptions at Mt. Erebus (here the magma inside the conduit could be directly observed by the radar) do not show a pulsed behavior. Pulsed behavior in this context means that a Strombolian eruption is composed out of one single explosion, in which the eruption velocity decays continuously until the eruptions ends. A pulsed eruption on the contrary is composed out of several single explosions where the eruption velocity decays and then suddenly increases again due to the next explosion and so on. Contrary to eruptions at Mt. Erebus eruptions at Stromboli and Yasur are dominated by pulsed behavior which is most likely either due to a) a train of bubbles consecutively exploding at the magma surface in the conduit or b) the variation of eruption velocities is due to a flow instability of the fragmented magma inside the conduit prior to the release of the magma at the vent. The widening of the conduit prior to an eruption at Stromboli volcano was measured to be on the order of 1-3 mm using the modified Doppler radar system. We think that this conduit widening is associated with the arrival of an over pressurized gas bubble at the magma/air interface in the conduit. The opening angle of single eruptions was found to be on the order of 40 to 60°. Furthermore during the first 300 ms of an eruption the opening was found to slightly decrease. This was interpreted to be related to lowering of the magma surface in the conduit during an eruption. One of the key observations was the pulsed behavior of eruptions at Stromboli and Yasur volcano. We first developed a semi-analytical solution for the the rise of a gas slug in a conduit filled with a magma exhibiting a non-newtonian viscosity. This solution is an essential prerequisite to test fully numerical CFD models for the rise of bubbles in a magma exhibiting non newtonian behavior. We found quite a satisfactory agreement between the semi-analytical and the fully numerical solution. In a next step we implemented a constriction into the conduit of our fully numerical model to explore the possibility of generating slug trains when a large gas slug passes through such a constriction. We found that slug trains can be generated but the interval of ”explosions” at the magma surface was much larger than the ones observed in natural systems. We are currently repeating the same calculations with a model that includes a magma with a non newtonian viscosity. In an attempt to compare the activity of Yasur and Stromboli volcano, we first generated an event catalogue of eruptions at Stromboli as well as Yasur from which we then derived a magnitude based catalogue. This was found to follow a power law distribution, meaning that the process behind the eruption is a single, scale invariant process. Next we determined inter event times which were ranked into classes and again this distribution follows a power law. In addition the power spectra of the inter event times were found to follow a 1/f distribution with a slope of -1 which suggests a non-random process underlying the activity at both volcanoes. In a final step we developed a conceptual model that could explain the non randomness of the process based on a percolation theory model.
