In general, an understanding of the complex processes acting before and during volcanic eruptions is approached from various different sides, e.g. laboratory experiments on fragmentation and/or bubble burst eruption mechanisms, petrological analysis of the eruptive products and various geophysical monitoring and source localization techniques. Each of these techniques can deliver valuable insights by adding pieces of information about the physical processes that drive the volcanic activity. However, often studies are focussing on a single aspect of the process, without setting the results in a more general context. Often, this strategy is absolutely valid, when the focus is laid on a single piece in the complex chain of processes taking place in volcanic eruptions. This must fail when the results

aim to suggest a valid model for the combined observations at volcanoes using the above described techniques. The resulting models of volcanic source mechanisms and eruptive

features can therefore lead to biased assumptions. This study aims to close this gap between laboratory experiments, petro-chemical analysis and modern geophysical monitoring and source localization techniques in a case study of Mt. Yasur (Vanuatu) volcano.

The presented laboratory experiments on explosive volcanic eruptions upon rapid decompression show that decompression rate is the dening parameter in the experiments and that

a scaling to large-scale processes is valid. Furthermore, a model is presented that correlates measured particle velocities to decompression rate and initial gas-overpressure. This model is used to estimate source volumes and overpressures at Volcan de Colima (Mexico) and Mt. Yasur (Vanuatu).

A petrographically and geochemically characterization of Mt. Yasurs eruptive products suggests a shallow magma-mingling process at both of Mt. Yasurs active craters, perhaps due to rejuvenation of material slumped from the crater walls into an open conduit system. A study on the time-reversal imaging technique and its ability to detect the details of finite rupture (or time-variant) processes shows that the limitations of TR imaging start where the source stops being point-localised with respect to the used wavelength.

Inversion of the source mechanisms of Strombolian explosions at Mt. Yasur are performed using a multi-parameter dataset consisting of seismic, acoustic and Doppler-radar data. Time-reversal imaging and moment tensor inversion are used to invert the source location of the seismic long-period (f < 1Hz) signals, which is supposed to refl

ect

fluid movement at depth. The source is located in the north-east of the crater region in a depth of several

hundred meters. Furthermore, the source volume of the radiated infrasound signals is estimated from fundamental resonance frequencies. The results showed that the maximum

particle velocity measured with the Doppler radar correlates nicely with the estimated source volumes lengths. The inverted seismic moment does not show any correlation with the estimated slug sizes, i.e. the slug size does not map in seismic moment. This is an

important information, as it states that a larger source volume does not necessarily produces a larger seismic moment.

From these combined results, a common feeder system for all active craters at Mt. Yasur is proposed. The differences in event recurrence rate at the three active craters are believed to be controlled by either the conduit geometry or variations in degassing or cooling rate.

Strombolian-type eruptions at Mt. Yasur are suggested to be due to the burst of gas slugs with lengths and overpressures comparable to volcanoes showing similar eruptive patterns.

The results illustrate the importance of combined studies that overcome the limitations of single disciplines. In this way, a more comprehensive view of volcanic eruptions and the associated observations is possible. Such a multi-disciplinary approach will contribute to a better understanding of volcanic pro