We imaged the pumice of the two explosive events using the SEM at the University of Orléans. The aim was then to be able to trace the plagioclase microlites manually , and the oxides and bubbles automatically in order to obtain the glass proportion. At the same time, we obtained the bulk water content of the pumice using an Elemental Analyzer. We then combined these results to obtain the water content of the glass. The first part of the model uses our textural analyses to obtain the initial pressures and porosi-ties of our samples. Then, the second part of the model determines the initial depths of the samples based on these parameters. 2. Eruptive Depth and Pressure Model We used the two-part eruptive depth and pressure model developed in Burgisser et al. (2010; 2011) for the 88 Vulcanian explosions that occurred at Soufrière Hills (Montserrat) in 1997. Temperature magma-chbr (°C) 950 ρ magma (kg,m-3) 2455 Pressure magma-chbr (MPa) 300 X H2O max (%wt) 7,38 Merapi volcano (Java, Indonesia) is one of the most dangerous volcanoes on Earth. It is known to produce lava domes that collapse in deadly pyroclastic flows because of gravity or auto-explosivity. Even though this volcano is mostly effusive, explosive eruptions occur in its history, most recently in 2010. Two major explosive events occurred that year, the first one on October 26, and the second one on November 05. They produced explosive columns that quickly collapsed in pumice-rich pyroclastic flows that ran up to 15.5 km from the summit. The 2010 eruption was very well observed, making this event a good candidate for investigating the effusive-explosive transition with a full suite of geochemical and geophysical data. The pumice of these two events was analyzed and the data used in a two-step computer model in order to investigate the pre-explosive conditions. The explosive-effusive transition often occurs at arc volcanoes. The parameters driving this transition (overpressure, eruptive volume,. . .) are known but the relative importance of each parameter remains unclear. This study demonstrates the primacy of overpressure in determining the eruptive mode of the volcano, with volume and other parameters contributing only to the magnitude rather than the character of the event.